Journal of cave and karst studies

Citation
Journal of cave and karst studies

Material Information

Title:
Journal of cave and karst studies
Series Title:
Journal of Cave & Karst Studies
Alternate Title:
Continues NSS bulletin (OCLC: 2087737)
Creator:
National Speleological Society
Publisher:
National Speleological Society
Publication Date:
Language:
English

Subjects

Subjects / Keywords:
Geology ( local )
Genre:
serial ( sobekcm )
Location:
United States

Notes

General Note:
Articles: http://dx.doi.org/10.4311/2011LSC0223 Cave Cricket Exit Counts: Environmental Influences and Duration of Surveys / Floyd W. Weckerly http://dx.doi.org/10.4311/2010MB0136R Involvement of Bacteria in the Origin of a Newly Described Speleothem in the Gypsum Cave of Grave Grubbo (Crotone, Italy) / Paola Cacchio, Claudia Ercole, Rosita Contento, Giorgio Cappuccio, Maria Preite Martinez, Maddalena Del Gallo, and Aldo Lepidi http://dx.doi.org/10.4311/2011AN0219 The Prehistoric Cave Art and Archaeology of Dunbar Cave, Montgomery County, Tennessee / Jan F. Simek, Sarah A. Blankenship, Alan Cressler, Joseph C. Douglas, Amy Wallace, Daniel Weinand, and Heather Welborn http://dx.doi.org/10.4311/2010EX0167R Candidate Cave Entrances on Mars / Glen E. Cushing http://dx.doi.org/10.4311/2011LSC0215 The First Subterranean Freshwater Planarians from North Africa, with an Analysis of Adenodactyl Structure in the Genus Dendrocoelum (Platyhelminthes, Tricladida, Dendrocoelidae) / Abdul Halim Harrath, Ronald Sluys, Adnen Ghlala and Saleh Alwasel http://dx.doi.org/10.4311/2010AN0148R1 Micro-Charcoal Abundances in Stream Sediments from Buckeye Creek Cave, West Virginia, USA / Gregory S. Springer, L. Nivanthi Mihindukulasooriya, D. Matthew White, and Harold D. Rowe http://dx.doi.org/10.4311/2010ES0178R Response of the Karst Phreatic Zone to Flood Events in a Major River (Bohemian Karst, Czech Republic) and its Implication for Cave Genesis / Helena Vysoká, Jir¡ í Bruthans, Karel Z¡ ák, and Jir¡ í Mls http://dx.doi.org/10.4311/2011jcks0193 A New Species of Nicoletiidae (Insecta: Zygentoma) from Kartchner Caverns State Park, Arizona / Luis Espinasa, Robert B. Pape, Alanna Henneberry, and Christopher Kinnear http://dx.doi.org/10.4311/2011es0222 Regionalization Based on Water Chemistry and Physicochemical Traits in the Ring of Cenotes, Yucatan, Mexico / Rosela Pérez-Ceballos, Julia Pacheco-Ávila, Jorge I. Euán-Ávila, and Héctor Hernández-Arana http://dx.doi.org/10.4311/2011jcks0197 Delineating Protection Areas for Caves Using Contamination Vulnerability Mapping Techniques: The Case of Herrerías Cave, Asturias, Spain / A.I. Marín, B. Andreo, M. Jiménez-Sánchez, M.J. Domínguez-Cuesta, and M. Meléndez-Asensio http://dx.doi.org/10.4311/2011JCKS0213 Microbiological Activities in Moonmilk Monitored Using Isothermal Microcalorimetry (Cave of Vers Chez Le Brandt, Neuchatel, Switzerland) / Olivier Braissant, Saskia Bindschedler, Alma U. Daniels, Eric P. Verrecchia, and Guillaume Cailleau http://dx.doi.org/10.4311/2011LSC0216 Importance of Karst Sinkholes in Preserving Relict, Mountain, and Wet-Woodland Plant Species under Sub-Mediterranean Climate: A Case Study from Southern Hungary / Zoltán Bátori, László Körmöczi, László Erdös, Márta Zalatnai, and János Csiky.
Restriction:
Open Access - Permission by Publisher
Original Version:
Vol. 74, no. 1 (2012)
General Note:
See Extended description for more information.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
K26-02061 ( USFLDC DOI )
k26.2061 ( USFLDC Handle )
12022 ( karstportal - original NodeID )
0146-9517 ( ISSN )

USFLDC Membership

Aggregations:
Added automatically
Karst Information Portal

Postcard Information

Format:
serial

Downloads

This item has the following downloads:


Full Text

PAGE 1

rf rfrntbbrbbfnffb rf bfnbfbfnbbrf frbbf rfntbtbtrttbbb rf bnrfbfnffrfbrb bbbnbnnnrnbbbbtbrf rffbrfbfnf bfrnbrf bbfffbfffbnfbfffnbbnbb brfbbnffnf rtbrnbbbnrf nrfnffnbfbntrtrfbf bbtbrrntttbrf bfbfbnbbffffbrnn fbnfbrf bnrtbn bnrf nnbfnbfbffbnrfbfbftf rnfnbntbbbbbntbbrf ffbffbrbfnnnffbbrbnfbfn nnrfrbbtbbbrf fbbnbfrfrbffbfbfn rffrfbff bbbbbrrntbbnbnrf nnfnbbtbbfnnfbrfrfb nfbbf nnbtnbnbnfbrrrrf bfnfbtnbbffbffbn bffrfbrfbbf tbtnnfnttbbbnn rntbn b bttn rf ff r r r r rf

PAGE 2

r rbbnttrnfbnf fbbnfftfbfn rb t ffnbnnbnbnb nfbtfbfbbbb bfbfnfnbbnb fnbbfnbbnf rbbnttrn nfbfbbfffffb bbnbffnfbfnbfbf ffbnffb r bbbfnnfbbfnbfbb nfbbrbffbb bff fbbb rbbfnb bnffbbffbnff fbbbfnfbbn ffffbf rfnfnbffbbbbfb fffffbfnbfbbbbf nffnnnbntbf nbbb rfbfnbbfbbbfbf bbfnnfffbnfbfnbffbf fbbnfnbf rbnbrnfffb bfb rbbbnb tbfbfbfffbf fffbfbnbn fffbnffbbbfnbbfb fbfbfbnfbbb ffnfnbnf fbbbnb bfbffbnfbbnf fnnbfffbn bbbffbb fnnfbfffbb nbbffbbbffnn fntbffnnbf nnfbbffbbf fbfnfbffbb fbnnfbrbfbbbbb fbf nbfffnbf bbbfbfntffb fbbnntbn bffbbfbfbbb bbbfffffbn bfbfbfnb fbff fnbbbfbfnbbnbbf nbbffffbfbbffb bfnbfnbnfffr frbfbbfbffnb nnfbffbfb ffbffbfbf b rbffbfbb rfnfffr ffnfff rbfbfbfbfnbf bbfnbfnnbnfb bfnbbbbfbfnnb bbffnbbbfnf fnbbf bffbnbfbbfnf nffbfnnbfnbbf fbffnbfnfbbff nbbbrbbfbf fnb rbbbffnfn bfnfffbfbfbfbff bffffbfb bfbfffbffbnfrbbfbf nfbrbfffbf rrbnfnf b rbfnbf bbbnfbbbf ffnnbb nfbnbbbfbb ffbnff fbb bfbfbnnbbbnfn bbfnffnfb fbb fbbbfbf nfbfffbbffbb bnfb bbbbbnn bbbfbfbff nbfbrfbfbfb bbnfff bbbffb rb bfbbbnffnbfnt bfbbnnbbbfb fffbfbbbbbf nffffffnnn r bbfnbfbbbfbfb bbbfnbfbb bbfbnfbbftf ffbfnbfbbb ffbnfbbbnfnnbfn nbbrftfnf nfbfnnbfnbfbb bfbnb fbfbbfb rrrbbfb bbfbbffnnbfnrbbnttrnrfbnffbbf fbfbfnfnbrfbffff fnfbbnff rntbrbbnfbfntfnfbnffffbn fnfbbrbbnttrnrfbfff rbbnttrnnbnnnrbfbfbn frbrbbnfrnfffbnn rbbfbfnfnbn bnnbffrffnbtbfb fff fbfrbbf b nfnfb bfbnbn ff fbr nf fnfff r r ftrbfb nbbnnr r r rr bbntfbf nbtff nfbr nbfnff fbbnfnn b ffbb fbfbb bbfbfbbfbff bfb btfbf rfnn fbfftnrff ff fbb bfbbtfbb ftfbb fftffbf fbfftnbrr nffff bfbb bbfbrbr fbfb r fbbfbnn bfbbtbf ffbf

PAGE 3

CAVECRICKETEXITCOUNTS:ENVIRONMENTAL INFLUENCESANDDURATIONOFSURVEYS F LOYD W.W ECKERLY 1 1 DepartmentofBiology,TexasStateUniversity–SanMarcos,SanMarcos,Texas,78666,USA,fw11@txstate.edu Abstract: Cavecricketabundanceisusedasanindicatorofintegrityofcave ecosystems.Onemeansofmonitoringcavecricketabundanceiscountingcricketsas theyemergefromcaveentrancesfortwohoursaftersunset.Theinfluenceofcloudcover, relativehumidity,andsurfacetemperatureoncountsisunknownandtheremightbefew cavecricketsthatemergeduringthefirsthourofthesurvey.Usingmixedeffectsmodels, Iassessedtheinfluenceoftheseenvironmentalvariablesonexitcountsandestimated whencavecricketsemergedwithinthetwo-hoursurveyperiod.Exit-countsurveyswere conductedinelevencavesoverfouryearsincentralTexas,andcavesweresurveyedupto fourtimesayearacrossthefourcalendarseasons.Cloudcover,relativehumidity,and temperatureinfluencedcounts,butthegreatestinfluencewasfromtemperature.Peaksin cavecricketcountsoccurred80to90minutesafterthestartofasurveyanddeclined thereafter.Cavecricketexitcountsurveysshouldrecordsurfacetemperature,cloud cover,andrelativehumidityatthestartofsurveyssothatcountscanbeadjustedfor theseenvironmentalinfluences.Also,surveyscanbeshortenedto1or1.5hoursin length. I NTRODUCTION Cavecrickets( Ceuthophilus )areoftenkeystonespecies incaveecosystems(Faganetal.,2007;Lavoieetal.,2007; Tayloretal.,2007b).Theseinsectsprovideinputsof allochthonousresourcestocavesthatareoftenenergy depauperate,theguanofromcavecricketssustains invertebratecommunities,andtherearepredatorsthat specializeontheeggslaidbycavecrickets(Tayloretal., 2005).Duetotheinordinateinfluenceofcavecricketson caveecosystems,thenumberofcavecrickets(hereafter abundance)thatemergefromandreturntocavesis integraltotheconservationoftroglobiticendangered speciesandspeciesofconcern(Tayloretal.,2007b). Onemeansofmonitoringabundanceandtemporal trendsincavecricketsiswithexit-countsurveys(Taylor, etal.,2007b).Beginningatsunsetoneveningswhenthe surfacetemperatureisatleast5 u C,twosurveyorscount cavecricketsfortwohoursastheyemergefromacave entrance.Environmentalconditionsthatmightinfluence cavecricketemergencearetheamountofmoonlightand thetemperatureandrelativehumidityonthesurface (Campbell,1976;Poulsonetal.,1995;Yoderetal.,2011). Surfaceactivityofcavecricketsisreportedtobelower whenthereismoremoonlight,duringcoolorhotnighttime temperatures,andwhenrelativehumidityislower.But howtheseenvironmentalvariablesactuallyaffectthe numberofcavecricketsthatemergefromcaveshasnot beenexamined.Furthermore,cavecricketsarerarelyactive andforagingonthesurfaceduringdaylighthours(Campbell,1976;Yoderetal.,2011).InCarlsbadCaverns NationalPark,Campbell(1976)notedthatasmanyas 50%ofthecavecricketsthatemergedonsummernights didsobetweenoneandtwohoursaftersunset.Because cavecricketemergenceoccursaftersunset,thenecessityof conductingcountsbeginningatsunsetisquestionable. Perhapssurveysshouldbeginahalfhourtoonehourafter sunset,whenmostcavecricketsemerge.Also,ifmostcave cricketsemergeonetotwohoursaftersunset,needthese surveysbeconductedfortwohours? Theabilityofasurveytechniquetoestimatepopulationsisbasedontheaccuracyofactualpopulation estimatesinthesettingswherethesurveytechniquewill beapplied.Estimatesobtainedfromsurveysmustbe comparedtoknownpopulations(e.g.,Weckerlyand Foster,2010).Forcavecrickets,itwouldbequite challengingtoknowtheactualpopulationincaves,and thus,conductarobustevaluationofthereliabilityofexit counts.Nevertheless,understandingwhatinfluences countsandwhentosurvey,sothatcountsareconducted whenanimalsaremostlikelytoemerge,canbeusedto ensurethatsurveysarestandardized.Ifcountsare conductedwhenmostanimalsemerge,thencounts obtainedmayprovidereliableinformationoncavecricket abundance. Thespecificobjectivesofthisstudyweretodetermineif andhowcloudcover,relativehumidity,andtemperature influenceexit-surveycounts,todeterminethemagnitudeof theinfluenceoftheseenvironmentalvariablesoncounts, andtoestimatewhencavecricketsemergedfromcave entranceswithinthe2-hoursurveyperiod. S URVEYS Cavecricketexit-countsurveyswereconductedat elevensmallcavesintheBalconesCanyonlandsPreserve, F.W.Weckerly–Cavecricketexitcounts:environmentalinfluencesanddurationofsurveys. JournalofCaveandKarstStudies, v.74, no.1,p.1–6.DOI:10.4311/2011LSC0223 JournalofCaveandKarstStudies, April2012 N 1

PAGE 4

TravisCounty,Texas,USA(Table1).TheBalcones CanyonlandsPreserveisadiscontinuouscollectionof propertiesconsistingof5,365hamanagedforendangered species.Thesurfaceenvironmentisamixofashejuniper ( Juniperusashei ),hardwoods(mostlyoaks, Quercus spp.), shrublands,andgrasslandswithgentlyslopingtosteepsidedcanyons.Daytimetemperaturesarehot(often 35 u C)insummerandmild(5to25 u C)inwinter(Watson etal.,2008). Exit-countsurveyswereconductedfrom2006to2010. Asurveyconsistedoftwosurveyorscountingallcave crickets(juveniles,nymphs,andadults)thatemergedfrom acaveopeningina2-hourperiodthatbeganatsunset.Red lightswereusedduringsurveystoaidinseeingcave cricketsandreducedisturbingthem.Speciesofcave cricketswere Ceuthophilussecretus ,sometimes C.cunicularis ,andaspeciesof Ceuthophilus yettobedescribed (speciesB,Tayloretal.,2007a).Atthebeginningofthe surveys,thepercentageofcloudcover(avisualestimate), therelativehumidity,andthetemperaturewererecorded. Percentagecloudcoverwasapracticalwaytoaccountfor someofthevariationinlightfromthemoonandstars. Countsurveyswereconductedinallfourcalendarseasons oftheyear.Ateachcavesixtotwelvesurveyswere conducted. A NALYSIS –E NVIRONMENTAL I NFLUENCES Ianalyzeddatausingmixed-effectsmodelsbecauseexitcountsurveyswererepeatedlyconductedateachcave acrossthefouryears(PinheiroandBates,2000).Because theresponsevariablewasacount,Iusedageneralized linearmixed-effectsmodelassumingaPoissondistribution (Faraway,2006). Fixedfactors(hereafter,variables)werecloudcover, temperature,andrelativehumidity.Variablesthatpotentiallyinfluencedtheresponse,countsofemergingcave crickets.Becausesurfaceactivityofcavecricketsshould increasefromcooltowarmtemperaturesbutdeclinewith hottemperatures,Iestimatedaquadraticrelationship betweentemperatureandcounts(Poulsonetal.,1995). Cloudcover,relativehumidity,andtemperaturewere continuousvariables(SokalandRohlf,1995).Mixed-effect modelscansufferfromcomputationalinstability.The iterativealgorithmsandoptimizationprogramsthat estimateparametersmaynotconverge,parametersare estimatedbutconfidenceintervalsarenot,ortheestimated confidenceintervalsareverywide(PinheiroandBates, 2000).Toreducethepossibilityofcomputationalinstability,Icenteredcontinuousvariables(Chengetal.,2010)by subtractingthevaluesfromthemeans. z -testsofcoefficientsofvariableswereusedtoassessifavariablehadan influenceoncounts(Faraway,2006). Toassessthemagnitudeofinfluence(MI)ofcloud cover,relativehumidity,andtemperature,Icalculated Y max { Y min jj = X max { X min jj ,wherethelargestpredicted countandits X valuewere Y max and X max andthesmallest predictedcountandits X valuewere Y min and X min .AMI withalargevalueindicatesthatthevariablehadalarge influenceoncounts.IcalculatedanMIinthismanner becauseitisnotpossibletoassessthemagnitudeof influenceofvariablesbycoefficientsor z -testswhenthe unitsofpredictorsarenotthesame(SokalandRohlf, 1995). Cavewastreatedasarandomfactor.Randomfactors werepartofrandomeffects,variancesthatallowrelationshipsbetweenfixedvariablesandtheresponsevariableto changeamongcaves(Chengetal.,2010).Inthegeneralized linearmixed-effectmodel,therewasonerandomeffect,the cave.Thisallowedinterceptsoftheregressionsbetween variablesandcountstovaryamongthecaves(Chengetal., 2010). Togaugethevariabilityintherelationshipsbetween variablesandcounts,Iestimatedanintra-classcorrelation Table1.Featuresandnumberofcavecricketexitcountsurveysconductedfrom2006–2010ontheBalconesCanyonlands Preserve,TravisCounty,Texas.Also,thenumberofsurveyswithadequatecounts( $ 20)foranalysesofsurveylengthis reported.Adashedlinedenotestherewerenosurveyswithdata. CaveLength,mDepth,mNo.ofsurveys No.ofsurveysfor analysisofsurveylength Amber10.27.5108 Cotterell7.97.61111 DistrictPark74.713.41111 FlintRidge283.046.51111 Gallifer35.07.41110 Kretschmarr7.66.11312 KretschmarrDP8.011.46… Rootcomplex15.74.377 Spider9.07.41212 Tardus6.06.37… Tooth50.55.696 C AVECRICKETEXITCOUNTS:ENVIRONMENTALINFLUENCESANDDURATIONOFSURVEYS 2 N JournalofCaveandKarstStudies, April2012

PAGE 5

coefficient.Thecoefficientiscalculatedbydividingthe varianceoftheinterceptrandomeffectbythesumofthe residualandinterceptvariances(PinheiroandBates,2000). Theresidualvariancemeasuresthesquareddeviation betweenobservedandpredictedresponsevariables,just likeinanordinaryleast-squaresregression.Thequantity canrangefrom0to1,with0indicatingnovariabilityin relationshipsacrosscavesand1denotingtremendous variationinrelationshipsacrosscaves. A NALYSIS –T WO H OUR S URVEY P ERIOD Mixed-effectsmodelswerealsousedtoestimate emergenceofcavecricketsinthe2-hoursurveyperiod. However,becausetheresponsevariablewascontinuous,I usedlinearmixedeffectsmodels.Theresponsevariablewas thecountduringa10-minuteintervaldividedbythe2-hour totalandexpressedasapercentage.Ilabeledtheresponse variablepercentcount.Fixedvariablesweretime,cloud cover,temperature,andrelativehumidity.Minuteswasa discretevariable,coded1to12toreflectthetwelve10minuteincrementsinthe2-hoursurveyperiod.The continuousvariablesofcloudcover,temperature,and relativehumiditywerecentered.Minutesweremodeledas acubicrelationshipwithpercentcounts;minuteshadlinear ( minutes ),quadratic( minutes 2 ),andcubic( minutes 3 ) coefficients.Acubicrelationshipallowedforthepossibility ofalowpercentcountearlyinthetwohoursurveyperiod andapeakandadeclineincounts,iftheyexisted,atthe endofthesurveyperiod.BecauseIusedalinearmixedeffectsmodel,Iconductedananalysisofvariancetoassess whichvariableswereinfluencingpercentcount(not possiblewithageneralizedlinearmixed-effectsmodel). InthisanalysisImodeledrandomeffectssothat relationshipsoffixedvariablescouldvaryin intercepts minutes ,and minutes 2 .Becauseofthenumberofrandom effects,Pearson’scorrelationcoefficientswerealsoestimatedtoaccountforpossiblecovariancebetweenrandom effects.Thelinearmixed-effectsmodelhadcorrelation coefficientsbetween intercepts and minutes intercepts and minutes 2 ,and minutes and minutes 2 (PinheiroandBates, 2000).Correlationcoefficientsestimatedbetweenrandom effectscanhavewideconfidenceintervals,indicativeof computationalinstability(PinheiroandBates,2000).In casethisoccurred,Imodeledtherandomeffectsassuming therewerenocorrelations(i.e., r 5 0.0).Apreliminary examinationofthedataindicatedtheassumptionof homoscedasticitywasviolated.Therefore,avariance functiontocharacterizeheteroscedasticitywasused(PinheiroandBates,2000).Thevariancefunctionallowed residualvariancetoincreasewithincreasingvaluesof minutes F INDINGS –E NVIRONMENTAL I NFLUENCES Therewere108exit-countsurveysconductedinthe elevencaves(Table1).Acrossthesesurveys,countsranged from1to1536(Table2).Thecloudcoverduringsurveys rangedfromopensky(nocloudcover)tocompletecloud cover,withamediancloudcoverof20percent.Relative humidityrangedwidely,buthalfthesurveyswere conductedwhenrelativehumiditywasbetween50and75 percent.Thefirstandthirdquartilesoftemperaturesatthe startofsurveyswere18 u Cand28 u C,respectively. Therewasaquadraticrelationshipbetweentemperatureandcounts(Table3).Cloudcoverandrelative Table2.Quartiles(first-25 th percentile,second-medianor50 th percentile,third-75 th percentile),minimum,andmaximum valuesofpercentcloudcover,percentrelativehumidity,temperatureinCelsius,andcountsduringthe108cavecricketexit countsurveysconductedin11cavesontheBalconesCanyonlandsPreserve,TravisCounty,Texas. PercentileCountCloudCoverRelativeHumidityTemperature Minimum10167 Firstquartile2905018 Median149206024 Thirdquartile258427528 Maximum15361009834 Table3.Coefficients,standarderrors,andfindingsfrom z -tests( z value, P value)forfixedvariablesandinterceptforthe generalized,linearmixedeffectsmodel.Temperatureandtemperature 2 estimatedaquadraticrelationship. CoefficientValueSE zP Intercept5.1690.246520.97 0.001 Cloudcover0.0030.000218.10 0.001 Relativehumidity0.0020.0004 2 8.96 0.001 Temperature0.0210.001120.35 0.001 Temperature 2 2 0.0010.00015.27 0.001 F.W.W ECKERLY JournalofCaveandKarstStudies, April2012 N 3

PAGE 6

humidityalsoinfluencedcounts.Countsincreasedfrom cooltemperatures(5–10 u C)towarmtemperatures( 25 u C)anddeclinedwithhottertemperatures(Fig.1).Both cloudcoverandrelativehumidityhadpositiverelationshipswithcounts. ToestimatetheMIfortemperature,Imultiplied regressioncoefficientsforcloudcoverandrelativehumiditybytheirrespectivemeans.ToestimateMIforcloud coverandrelativehumidity,Ilikewisemultipliedtheother coefficientsbytheirmeans.TheMIfortemperature,cloud cover,andrelativehumiditywere9.2,0.20,and0.13, respectively. Theresidualvarianceforthegeneralizedlinearmixedeffectsmodelwas0.76,andtheintra-classcorrelation coefficientwas0.43.Therelationshipsofcloudcover, relativehumidity,andtemperaturewithcavecricketexit countsdidvaryacrosscaves,butthedifferenceswerenot dramatic. I NTERPRETATION –E NVIRONMENTAL I NFLUENCES Allthreeenvironmentalvariablesinfluencedcountsof cavecricketsemergingfromcaves.Buttemperatureclearly hadamuchgreaterinfluenceoncountsthandidcloud coverorrelativehumidity.TheMIfortemperature indicatedthatcountsincreasedover900percentforevery onedegreeincreaseintemperaturebetween5and25(i.e., countsincreased,onaverage,byninecavecricketsfor everyonedegreeincreaseintemperature),whereascloud coverandrelativehumidityincreasedbyonly13to20 percentforeveryonepercentincreaseinthesevariables. Theinfluenceoftemperatureoncounts,however,wasnot astraightforwardlinearincrease.Accordingtomy analysis,cavecricketsaremostactiveintheBalcones CanyonlandsPreservewhennighttimesurfacetemperaturesarebetween20and30 u C.Thereductionincave cricketemergenceoncoolandhotnightsisconsistentwith studiesindicatinghottemperaturesarestressfulforcave cricketsandthatcavecricketactivityiscurtailedwhen surfacetemperaturesarecool(Poulsonetal.,1995;Studier andLavoie,1990). Anotherfindingfromtheanalysiswasthatrelationshipsbetweenenvironmentalinfluencesoncountsvaried somewhatacrosscaves.Therelationshipsvariedin interceptsbutnotinslopes.Therefore,theformof relationshipsdidnotchangeacrosscaves.Forexample, thequadraticrelationshipbetweentemperatureandcounts wasthesamefromcavetocave.Thevariabilityin interceptsacrosscavessuggeststhatcavesvariedincave cricketabundance. F INDINGS –S URVEY L ENGTH FortheanalysisofpercentcountsIonlyusedsurveys whereaminimumoftwentycavecricketswerecounted, consequentlytherewasdatafromninecavesand88exitcountsurveys(Table1).Intheinitialmodeltherewereno influencesonpercentcountfromtemperature,relative humidity,orcloudcover( F 1,1038 1.32, p 0.179). Anothermodel(labeledfullrandomeffectsinTable4)was analyzedwiththesevariablesremoved.Thevariablesinthe fullrandomeffectsmodelwere minutes minutes 2 ,and minutes 3 .Thefullrandomeffectsmodelhadissueswith computationalinstability.Twoofthethreecorrelation coefficientshadenormouslywideconfidenceintervals,and Iwasnotconfidentthatthismodelestimatedthese parametersreliably.Thus,Iranthefinalmodel(subsets randomeffectsinTable4)withoutthecorrelationcoefficientsbetweenallpairsofrandomeffects(Table4). Confidenceboundsonestimatesofallparameters(coefficients)andrandomeffectsofthesubsetsrandomeffects modelindicatedthatcomputationalinstabilitywaslessofa problem.Thethreecoefficientsfor minutes (linear,quadratic,cubic)wereusefultomodelingtherelationshipbetween minutes andpercentcountbecauseeveryconfidence Figure1.Graphsshowingrelationships(and1standard errorenvelopes)betweenrelativehumidity,cloudcover, temperature,andcounts.Notethatthescaleforcountson theYaxisdiffersbetweenthetemperaturegraphandthe othertwographs. C AVECRICKETEXITCOUNTS:ENVIRONMENTALINFLUENCESANDDURATIONOFSURVEYS 4 N JournalofCaveandKarstStudies, April2012

PAGE 7

intervaldidnotoverlapzero.Also,itwasagoodideato modeltheheteroscedasticityasincreasingresidualstandarddeviationwithincreasingpercentcountsbecausethe variancefunctionhadconfidenceintervalsthatdidnot overlapzero.Theintra-classcorrelationcoefficientwas 0.38forthesubsetsrandomeffectsmodel.Thecubic relationshipbetween minutes andpercentcountdidvary amongcaves,butthevariationwasmodest. Thepeakincountsoccurredbetween80and90minutes intotheexit-countsurveys(Figure2).Countsincreased dramaticallyfromthebeginningofthesurveytothetime ofthepeakanddeclinedthereafter.Iusedthepredicted percentcountateach10-minuteincrementtocalculatethe cumulativepercentcount.Duringthelast1.5hoursofa survey,93percentofallcavecricketswerecounted,and duringthelast1.0hoursofacountsurvey,69percentofall cavecricketswerecounted. Thedeclineinpercentcountsinthelasthalfhourof surveysindicatesthatextendingsurveysbeyond2hours aftersunsetwillnotresultincontinuedhighcounts,a findingthatisconsistentwithemergencepatternsofcave cricketsinCarlsbadCavernsNationalPark(Campbell, 1976).Mostcricketsinthatstudyemergedonetotwo hoursaftersunset.Itwasnotsurprisingthatcloudcover, relativehumidity,andtemperaturedidnotinfluencetime historyofthecountsbutdidinfluencetotalcounts. C ONCLUSIONS Proactiveconservationofendangeredspeciesand speciesofconcernincaveecosystemsrequiresmonitoring ofspeciesthatindicatetheintegrityofthecaveecosystem (Lavoieetal.,2007;Tayloretal.,2007b).IncentralTexas, Ceuthophilus cavecricketsaretheindicatorspeciesofcave productivityinthekarstenvironment(Tayloretal., Table4.Estimatesand95percentlowerandupperboundsoffixedandrandomeffectsofthemixedeffectsmodelsthat estimatedrelationshipbetweenminutesafterstartofcavecricketexitcountsurvey(minutes)andthepercentageofallcave cricketsthatwerecountedin10minuteincrements.Thefullrandomeffectsmodelincludedallcorrelationsbetweenrandom components.Becausethecorrelationestimateshadwideconfidenceintervalsasubsetsrandomeffectsmodelwasestimated withoutcorrelationestimates.Minutes 2 isaquadraticcoefficientandminutes 3 isacubiccoefficient.Standarddeviationis denotedbySandRdenotesPearsonÂ’scorrelationcoefficient. Fixed Parameter Lower BoundCoefficient Upper Bound Random Effect Lower BoundEstimate Upper Bound FullRandomEffects Intercept9.0310.6912.36S(residual)2.342.663.01 Minutes0.811.422.04S(intercept)1.392.384.08 Minutes 2 2 0.33 2 0.20 2 0.06S(minutes)0.510.851.43 Minutes 3 2 0.03 2 0.02 2 0.01S(minutes 2 )0.120.200.33 Â…Â…Â…Â…R(intercept,minutes) 2 0.99 2 0.940.90 Â…Â…Â…Â…R(intercept,minutes 2 ) 2 1.00 2 0.990.73 Â…Â…Â…Â…R(minutes,minutes 2 )0.470.940.99 Â…Â…Â…Â…Variancefunction0.400.470.54 SubsetsRandomEffects Intercept9.2010.7312.25S(residual)2.412.743.10 Minutes0.811.412.02S(intercept)1.232.153.76 Minutes 2 2 0.32 2 0.20 2 0.08S(minutes)0.510.831.36 Minutes 3 2 0.03 2 0.02 2 0.01S(minutes 2 )0.110.180.30 Variancefunction0.380.450.52 Figure2.Therelationship,and1standarderrorenvelope, betweenminutesafterthestartofacavecricketexit-count surveyandthepercentageofallcavecricketsthatwere countedinthepreceding10-minuteinterval. F.W.W ECKERLY JournalofCaveandKarstStudies, April2012 N 5

PAGE 8

2007a).Informationgatheredfrommonitoringcavecricket abundanceisusedtoestimatechangesovertimeinthe numberofcavecricketsthatemergefromcaveentrances, determinethestateofcaveecosystems,assesswhether goalsofaconservationprogramaremet,andimplement actionstomitigatedegradationofthecaveenvironment. Fundamentaltomonitoringcaveecosystemsisthe protocolusedtocollectdataoncavecricketabundance. Astandardizedsurveyprotocolismerelyasetof instructionsthatdictatethatsufficientdataarecollected thatanalysesofthedatacanbedonewithstatistical methodsthatcanaccountforpeculiaritiesinthedataand heterogeneityinenvironmentalconditionsthatinfluence cavecricketemergence.Standardizationinhowthedatais collectedisfundamentaltoassessingtrendsacrosstime,a featurealreadypresentincavecricketexit-countsurveys.A featureofthesedatawasthatcountsurveyswere conductedperiodicallyacrosstimeinthesamecaves, whichrequiredtheuseofmixed-effectsmodelsthatcan accommodatethegroupingofdatabycaves(Pinheiroand Bates,2000).Itisclearthatheterogeneityincountscanbe duetotemperature,cloudcover,andrelativehumidity.Of theseenvironmentalinfluencestemperaturehasthegreatestinfluenceoncavecricketemergence.Toaccommodate environmentalheterogeneityinsurveystherearetwo options.Thefirstistorestrictsurveystowhenenvironmentalconditionsaremostamenabletoemergence.A drawbacktothisoptionisitwillconstrainthetimeswhen surveyscanbeconducted.Thefindingsofmyanalysis suggestthat,attheveryleast,surveysshouldbeconducted whensurfacetemperaturesarebetween20and30 u C.The otheroptionistoretaintheexistingprotocol.Byrecording temperature,cloudcover,andrelativehumidityatthestart ofthesurvey,theinfluenceofthesevariablesonexitcounts canbeaccountedforinanalysestodetecttemporalor spatialtrendsincavecricketabundance. Incavecricketexit-countsurveys,oneaspectofthe requiredsurveyeffortisthelengthoftimetoconduct survey.(Theotherisconductingcountsatcaveentrances whencavecricketsemerge.)Becauseoflogisticaland monetaryconstraints,monitoringprogramsrarelyenumeratetheentiretargetedpopulation.Instead,asubsetof thetargetedpopulationiscounted.Ifthecountistoreflect theactualpopulationstate,surveysshouldbeconducted whenindividualsinthepopulationaremostreadily counted.ForsurveysontheBalconesCanyonlands Preserve,thatwindowoftimeappearstobe1to1.5hours beginningahalfhourtoonehouraftersunset. A CKNOWLEDGEMENTS IthankJasonHuntforcollatingthedataandconducting theinitialanalyses.BenTobingenerouslyprovidedliterature, andBenjaminSchwartzreadanearlierdraftandoffered manygoodsuggestions.Ialsothankthemanybiologistsand staffwiththecityofAustinandTravisCountyTexasfor collectingthedataandtheirassistanceandfeedback.The projectwasfundedbythecityofAustin. R EFERENCES Campbell,G.D.,1976,Activityrhythminthecavecricket, Ceuthophilus conicaudus Hubbell:AmericanMidlandNaturalist,v.96,p.350–366. Cheng,J.,Edwards,L.J.,Maldonado-Molina,M.M.,Komro,K.A.,and Muller,K.E.,2010,Reallongitudinaldataanalysisforrealpeople: buildingagoodenoughmixedmodel:StatisticsinMedicine,v.29, p.504–520.doi:10.1002/sim.3775. Fagan,W.F.,Lutscher,F.,andSchneider,K.,2007,Populationand communityconsequencesofspatialsubsidiesderivedfromcentralplaceforaging:TheAmericanNaturalist,v.170,p.902–915. doi:10.1086/522836. Faraway,J.J.,2006,ExtendingtheLinearModelwithR:Generalized Linear,MixedEffectsandNonparametricRegressionModels,New York,Chapman&Hall,CRCTextsinStatisticalScience,312p. Lavoie,K.H.,Helf,K.L.,andPoulson,T.L.,2007,Thebiologyand ecologyofNorthAmericancavecrickets:JournalofCaveandKarst Studies,v.69,p.114–134. Pinheiro,J.C.,andBates,D.M.,2000,Mixed-EffectsModelsinSandSPLUS,NewYork,SpringerVerlag,548p. Poulson,T.L.,Lavoie,K.H.,andHelf,K.L.,1995,Long-termeffectsof weatheronthecricket( Hadenoecussubterraneus ,Orthoptera, Rhaphidophoridae)guanocommunityinMammothCaveNational Park:AmericanMidlandNaturalist,v.134,p.226–236. Sokal,R.R.,andRohlf,F.J.,1995,Biometry:ThePrinciplesandPractice ofStatisticsinBiologicalResearch,3rdedition,NewYork,W.H. FreemanandCompany,887p. Studier,E.H.,andLavoie,K.H.,1990,Biologyofcavecrickets, Hadenoecussubterraneus ,andcamelcrickets, Ceuthophilusstygius (Insecta:Orthoptera):metabolismandwatereconomiesrelatedtosize andtemperature:ComparativeBiochemistryandPhysiology,PartA, v.95,p.157–161.doi:10.1016/0300-9629(90)90025-N. Taylor,S.J.,Krejca,J.K.,andDenight,M.L.,2005,Foragingrangeand habitatuseof Ceuthophilussecretus (Orthoptera:Phaphidophoridae), akeytrogloxeneincentralTexascavecommunities:American MidlandNaturalist,v.154,p.97–114.doi:10.1674/0003-0031(2005) 154[0097:FRAHUO]2.0.CO;2. Taylor,S.J.,Krejca,J.K.,andHackley,K.C.,2007a,ExaminingPossible ForagingDifferencesinUrbanandRuralCaveCricketPopulations: CarbonandNitrogenIsotopeRatios( d 13 C, d 15 N)asIndicatorsof TrophicLevel:IllinoisNaturalHistorySurveyTechnicalReport 2007(59),97p. Taylor,S.J.,Weckstein,J.D.,Takiya,D.M.,Krejca,J.K.,Murdoch,J.D., Veni,G.,Johnson,K.P.,andReddell,J.R.,2007b,Phylogeographyof CaveCrickets( Ceuthophilus spp.)inCentralTexas:AKeystone TaxonfortheConservationandManagementofFederallyListed EndangeredCaveArthropods:IllinoisNaturalHistorySurvey TechnicalReport2007(58),45p. Watson,C.A.,Weckerly,F.W.,Hatfield,J.S.,Farquhar,C.C.,and Williamson,P.S.,2008,Presence-nonpresencesurveysofgoldencheekedwarblers:detection,occupancyandsurveyeffort:Animal Conservation,v.11,p.484–492.doi:10.1111/j.1469-1795.2008.00204.x. Weckerly,F.W.,andFoster,J.A.,2010,BlindcountsurveysofwhitetaileddeerandpopulationestimatesusingBowden’sestimators: JournalofWildlifeManagement,v.74,p.1367–1377.doi:10.1111/ j.1937-2817.2010.tb01259.x. Yoder,J.A.,Benoit,J.B.,LaCagnin,M.J.,andHobbs,H.H.,2011, Increasedcavedwellingreducestheabilityofcavecricketstoresist dehydration:JournalofComparativePhysiologyB:Biochemical, Systemic,andEnvironmentalPhysiology,v.181,p.595–601. doi:10.1007/s00360-011-0555-5. C AVECRICKETEXITCOUNTS:ENVIRONMENTALINFLUENCESANDDURATIONOFSURVEYS 6 N JournalofCaveandKarstStudies, April2012

PAGE 9

INVOLVEMENTOFBACTERIAINTHEORIGINOFA NEWLYDESCRIBEDSPELEOTHEMINTHEGYPSUM CAVEOFGRAVEGRUBBO(CROTONE,ITALY) P AOLA C ACCHIO 1 ,C LAUDIA E RCOLE 1 ,R OSITA C ONTENTO 1 ,G IORGIO C APPUCCIO 2 M ARIA P REITE M ARTINEZ 3 ,M ADDALENA D EL G ALLO 1 AND A LDO L EPIDI 1 Abstract: Microorganismshavebeenshowntobeimportantactiveandpassive promotersofredoxreactionsthatinfluencetheprecipitationofvariousminerals, includingcalcite.Manytypesofsecondarymineralsthoughttobeofpurelyinorganic originarecurrentlybeingreevaluated,andmicrobialinvolvementhasbeendemonstrated intheformationofpoolfingers,stalactitesandstalagmites,cavepisoliths,and moonmilk.Westudiedthepossibleinvolvementofbacteriaintheformationofanew typeofspeleothemfromGraveGrubboCave,thethird-largestgypsumcaveinItaly. Thespeleothemwestudiedconsistedofalargeaggregateofcalcitetubeshavinga complexmorphology,reflectingitspossibleorganicorigin.Weisolatedanabundant heterotrophicmicrofloraassociatedwiththisconcretionandidentified Bacillus Burkholderia ,and Pasteurella spp.amongtheisolates.AlloftheisolatesprecipitatedCaCO 3 invitrointheformofcalcite.Onlyoneoftheisolatessolubilizedcarbonate.Therelative abundanceofeachisolatewasfoundtobedirectlyrelatedtoitsabilitytoprecipitate CaCO 3 atcavetemperature.Wesuggestthathypogeanenvironmentsselectformicrobes exhibitingcalcifyingactivity.Isotopicanalysisproducedspeleothem d 13 Cvaluesofabout –5.00 % ,confirmingitsorganicorigin.ThelightestcarbonatespurifiedfromB4Magar plateswereproducedbythemostabundantisolates.SEManalysisofthespeleothem showedtracesofcalcifiedfilamentousbacteriainteractingwiththesubstrate.Spherical biolithspredominatedamongtheonesproducedinvitro.Withinthecrystalsproducedin vitro,weobservedbacterialimprints,sometimesinapreferredorientation,suggesting theinvolvementofaquorum-sensingsysteminthecalcium-carbonateprecipitation process. I NTRODUCTION FewgypsumkarstsystemsinItalyhavebeenstudiedin detail,butgeneralfeaturesthatdistinguishthesesystems fromlimestonesystemshavenonethelessbeenidentified. Thesekarstsystemsnormallycontainsmallphysicaland chemicaldepositswithverysimilarmorphologiesand chemicalcompositions,buttheymayalsocontainunusual speleothemsandcavemineralscharacteristicofgypsum caves. GraveGrubboCave(Crotone,SouthernItaly)contains twounusualcalcitedeposits(FortiandLombardo,1998). Thefirstisafloatingcalcitespeleothemresemblingagroup ofhalf-bubbles.Thesecondisanewlydiscoveredtypeof speleothemthatwassuspectedtobeoforganicorigindue toitscomplexstructurecomposedofalargeaggregateof iso-orientedcalcitetubes.PoluzziandMinguzzi(1998) havesuggestedthattheoriginsofthisspeleothemmaylie intheactivitiesofatroglobiticinsectthatbelongstothe order Trichoptera (genus Wormaldia? )anddwellsinthe sulfurcompound-richhypogeanwater.Intheory,these insectswouldbethrophicallysupportedbynutrientssynthesizedbychemoautotrophicbacteria.Inahypogean environment,thepresenceofreducedchemicalcompounds, suchashydrogensulfide,createsaredoxinterfacebetween thesecompoundsandtheoxygenfromthecaveatmosphere orfromoxygen-richwaters.Chemoautotrophicmicroorganismscanliveatthisinterfacebyderivingenergyfrom theoxidationofreducedchemicalcompounds.Specialized sulphur-oxidizingmicrobialcommunities,inparticular,use thisenergytoproduceorganicmatterinsitu.Organic matter,inturn,canserveasthebasisforafoodchainthat cansustaininvertebratecommunitiesinhabitingthedeep recessesofhypogeniccaves(Fortietal.,2002).Ontheother hand,thehypogeanenvironmentalsohostsbacteriathatuse orproducelargeamountsofCO 2 duetotheirmetabolic activities.Accordingtothistheory,thisCO 2 reactswiththe gypsum-saturatedwaterinthepresenceofCa 2 + ,resultingin thedepositionofcalcitearoundtheinsectlarvae.However, thisdepositionhypothesisdoesnotaccountfortheabsence ofinsectsinthespeleothembody,thediameterofthecalcite *CorrespondingAuthor:paolacacchio@yahoo.it 1 DepartmentofBasicandAppliedBiology,MicrobiologyLaboratory,Universityof L’Aquila,Coppito,67010L’Aquila,Italy 2 CNR-InstituteofStructuralChemistry,POB10,00016MonterotondoSc.,Rome, Italy 3 DepartmentofEarthScience-LaSapienzaUniversityandIGAG-CNR,Rome, Italy P.Cacchio,C.Ercole,R.Contento,G.Cappuccio,M.P.Martinez,M.Del.GalloandA.Lepidi–Involvementofbacteriaintheorigin ofanewlydescribedspeleotheminthegypsumcaveofGraveGrubbo(Crotone,Italy). JournalofCaveandKarstStudies, v.74,no.1, p.7–18.DOI:10.4311/2010MB0136R JournalofCaveandKarstStudies, April2012 N 7

PAGE 10

tubesbeinggreaterthanthatofthelarvae,andthelackof evidencethatcontinentalTrichopteracanaccumulate lithoidelements(PoluzziandMinguzzi,1998). Manystudieshavedemonstratedtheeffectsofheterotrophicbacteria,actinomycetesinparticular,onthe formationofspecificspeleothems(Baskaretal.,2006, 2009;Can averasetal.,1999,2001;Grothetal.,1999;Laiz etal.,1999,2000;LeMe tayer-Levreletal.,1997;Melimet al.,2001;Northupetal.,2000).Bacteriacanprecipitate extracellularcalciumcarbonatethroughavarietyof processes(BartonandNorthup,2007;Ehrlich,2002; Riding,2000;SimkissandWilbur,1989).First,mineralizationoccursasabyproductofmicrobialmetabolismthat involveseitherautotrophicpathways,whichcandeplete CO 2 locally,orheterotrophicpathways,whichreleaseinto theenvironment(Castanieretal.,1999,2000).Inthese passivemechanisms,reactionssuchastheenzymatic hydrolysisofurea,andtheammonificationofaminoacids causeariseinpH,which,iffreeCa 2 + ispresent,resultsin theprecipitationofcalciumcarbonate. Second,carbonatenucleationtakesplaceonbacterial cellwallsduetoionexchange(i.e.,anactiveprocess) throughthecellmembranebypoorlyunderstoodmechanisms(Castanieretal.,1999).Itmayalsooccurthrough theadsorptionofdivalentcationstospecificnegatively chargedfunctionalgroupsonthecellwall(Rivadeneyra etal.,2000).Nevertheless,oncecarbonatehasbegunto precipitateonthecellsurface,thecellactsasanucleation site.Thisstageiscriticalininitiatingmineralprecipitation. SubsequentCaCO 3 precipitationmaybepurelyinorganic. Thethirdpossibilityinvolvesspecificproteinspresent intheextracellularpolymericsubstances(EPS)produced bythebacterialcommunitiesthatcantrapsedimentand areoftenessentialforadirectdepositionofmicrobial carbonate(Ercoleetal.,2007;Riding,2000). Bygrowingmicroorganismsinthelaboratoryunder controlledconditions,itispossibletodelineatetheirability toalterthechemistryoftheirmicroenvironmentand producebiominerals.Forthisreason,weusedlaboratorybasedculturestoinvestigatetheextenttowhichheterotrophiccalcifyingbacteriaareinvolvedintheformationof thespeleothemconstructionrecentlydiscoveredinGrave GrubboCave.Weisolatedandcharacterizedculturable heterotrophicmicroflorafromthesampleandassessed eachisolateforinvitrocalcificationandcarbonate solubilizationatvarioustemperatures.Wealsostudied themineralandisotopiccompositionsandthemorphology ofthecarbonateconcretionandofthepurifiedbioliths obtainedinlaboratorycultures. M ATERIALSAND M ETHODS TheGraveGrubboCaveopensat158mabovesealevel inthegypsumkarstofVerzino,ontheCalabrianIonian side,justatthefootoftheSilanMassif(Crotone,Southern Italy;Fig.1).Severaldeepkarstlandformswererecently discoveredinthisarea,andtheGraveGrubbo-Vallone Cufalosystemisthemaincaveidentified(Galdenziand Menichetti,1998).GraveGrubboisthethird-largest gypsumcaveinItalyandextendstoover2km.Three galleriesbranchofffromthecaveentrance,theRamodi Cenerentola,theGalleriaquadrata,andtheRamodel fiume(Fig.2;FortiandLombardo,1998).Anundergroundriver,thewaterofwhichcontainssulfateand calciumions,flowsthroughthelowersectionofthemain passage,theRamodelfiume,andcanbefollowedfor about1.5km.Thetemperatureofthecaveisabout15 u C. Therockformations,structuralsettlementofthekarst formationsinthisarea,andthehydrodynamicsofthis systemarenotconsistentwiththedevelopmentoflarge, widespreadchemicaldeposits.Indeed,calcitespeleothems areveryrareandsmall,andtherearealmostnogypsum speleothems. Westudiedanunusualandlikelybiogenicspeleothem intheGraveGrubboCave.Thesamplelocationand sampleareshowninFigures2and3,respectively. ThespeleothemsinGraveGrubboCavepresentas sectionalmassiveoracicularstructures,soitwasclear, eventothefirstcaverswhovisitedthecave,thatthis speleothemwasremarkablydifferentfromotherformations.FortiandLombardo(1998)werethefirsttonotethe peculiarcharactersofthedepositanddescribeditfora multidisciplinaryresearchprogramontheVerzinoarea (Ferrini,1998).Thepresenceofthiskindofspeleothem insideGraveGrubboislimitedtoaconfinedareanexttoa smallwaterfallandtheterminallake,ataboutthree quartersofthetotallengthofthecave. Thestudiedspeleothem,madeofcalcitetubesconnectedtogether,mantledalimitedevaporiticoutcrop.Froma lithologicalpointofview,thedepositisdividedintotwo separatelevels,eachhavingadifferentdirectionforits tubes’axesthatlikelyresultedfromtwodifferentflow directionsduringitsperiodofdeposition.Atpresent,the speleothemhasnoconnectionwiththeactualwaterflow andappearstobefossilized. Asamplewascollectedfromtheupperlayerofthe speleothemusingasterilechiselandsamplebag,andwas storedat4 u Cfor24huntilmicrobiologicalanalysis.From thissample,weobtainedtwosub-samplesundersterile conditions.Onecontainedmaterialfromtheexterior surfaceofthespeleothem(the‘‘outer’’subsample)and theothercontainedmaterialfromthespeleothem’sinterior (the‘‘inner’’subsample). Toisolatetheculturableheterotrophicbacteriaassociatedwiththeouterandinnersub-samplesfromtheupper layerofthespeleothem,weground1gofeachsampletoa powderusingasterilemortarandpestleandsuspendedthis powderinsalinesolution(0.9%NaCl).Serialdilutions werespreadintriplicateonB-4agar(Boquetetal.,1973) containg2.5gcalciumacetate,4.0gyeastextract,10.0g glucose,and18.0gBiolifeagarperliterofdistilledwater. ThepHwasadjustedto8.0usingNaOH.B-4mediumisan I NVOLVEMENTOFBACTERIAINTHEORIGINOFANEWLYDESCRIBEDSPELEOTHEMINTHEGYPSUMCAVEOF G RAVE G RUBBO (C ROTONE ,I TALY ) 8 N JournalofCaveandKarstStudies, April2012

PAGE 11

enrichmentmediumforcalcifyingbacteria;theacetateion isanenergysource,whiletheCa 2 + cationisusedbythe calcifyingbacteriatoprecipitatecalciumcarbonate.The presenceofglucosespeedstheprocess.Inoculatedplates wereincubatedat32 u Cfortwoweeks.Previousstudies demonstratedthatcoloniesfromcavesamplesgrowvery slowlyatcavetemperatureandthatthediversityofthe culturablegenerawassimilarwhetherthebacteriawere grownatcavetemperature(13 u C)oratlaboratory temperature(28 u C)(Grothetal.,2001;Laizetal., 2003).Individualcolonieswereselectedandpurifiedby streakingonB-4agar.Therelativeabundanceofeach isolate,withrespecttothetotalculturablebacterialmicroflora,wasdeterminedbydirectcountsonB-4agarplates. Purecalcifyingisolateswerestoredinliquidnitrogen ( 2 196 u C). Calcifyingisolateswerecharacterizedusingmorphophysiologicalandbiochemicalmethods.Cellandaggregate morphologywasstudiedunderalightmicroscope.GramstainingwasperformedwiththeColorGram2kit(bioMe rieux,Marcy-lÂ’Etoile,France). Allphysiologicaltestsinwhichtemperaturewasnot astudiedvariablewerecarriedoutat15 u C.OxygenrequirementwasstudiedbyincubatingisolatesonB-4agarin ananaerobicchamber.Oxidaseactivitywasassessedusing asolutionofN,N-dimethylp-phenylenediamineoxalate, ascorbicacid,and b -naphthol(Oxoid).Catalaseproductionwasdemonstratedonslidesbytheformationofbubbleswhenasuspensionoftheorganismtobetestedwas mixedwithadropof3%(v/v)hydrogenperoxide.Acid productionwasdeterminedbyAPI50CH,APISTAPH, API20EandAPI20NEteststrips(bioMe rieux)according tothemanufacturerÂ’sinstructions.Ureaseactivity,nitrate reduction,argininedihydrolase,gelatinase, b -galactosidase, lysinedecarboxylase,ornithinedecarboxylase,andtryptophandeaminaseactivitiesandH 2 Sindoleandacetoin productionwerealsocheckedbyAPIteststrips(bioMe rieux).APILABPlussoftware(updatev.3.3.3)wasused fortheAPItestcultureidentification. WeassessedthecalciteproductionofisolatesbyculturingthemonmediumB-4,asdescribedbyBoquetetal. (1973).Thebacterialisolateswerespreadintriplicateon Figure1.GeographicallocationofGrottaGraveGrubbo(GraveGrubboCave),Calabria,SouthernItaly. P.C ACCHIO ,C.E RCOLE ,R.C ONTENTO ,G.C APPUCCIO ,M.P.M ARTINEZ ,M.D EL .G ALLOAND A.L EPIDI JournalofCaveandKarstStudies, April2012 N 9

PAGE 12

thesurfaceofagarplates,whichwerethenincubatedinthe presenceofairat4,15,22,and32 u C.Weexaminedplates periodicallyunderalightmicroscopetomonitorcrystal productionforupto40daysafterinoculation.Fornegativecontrols,wecheckedforthepresenceofcrystalsin sterilemediumandmediuminoculatedwithautoclaved bacteria. Crystalsproducedbyculturedbacteriawereremoved fromthemediumbycuttingoutagarblocksandplacing theminboilingwateruntiltheagardissolved.Thesupernatantsweredecanted,andthesedimentwasresuspended andwashedindistilledwateruntilthecrystalswerefreeof impurities(Rivadeneyraetal.,1998).Thewashedcrystals wereair-driedat37 u CandthenusedtodetermineCaCO 3 yield,crystal-phase,stablecarbonandoxygenisotopecompositions,andmorphology. Thecarbonateyieldofthecrystalsobtainedafterthree monthsofincubationwasexpressedasapercentageofthe theoreticaldryweightofcalciumcarbonatebydividingthe experimentaldryweightofthecrystalsbythetheoretical dryweight.Thetheoreticaldryweightofcalciumcarbonatewasthestoichiometricquantityobtainedfromthe 18mgCa 2 + containedineachplate.Theexperimentaldry weightofCaCO 3 wasthemeanofthevaluesobtainedfrom fourindependentexperiments.Datawereanalyzedusing theStudentÂ’s t -test. X-raydiffraction(XRD)measurementsweredonebya two-circlet/2tdiffractometerwithaCuradiationsource, secondarygraphitemonochromator,andscintillationdetector(SeifertMZIV).ThesupplyvoltageoftheX-ray Figure3.Macroscopicmorphologyofthenewlydescribed speleothemfromGraveGrubboCaveshowingthetubeswith meanofinnerdiameter4.24mm(PoluzziandMinguzzi, 1998). Figure2.MapofGraveGrubboCave(FortiandLombardo,1998).SamplingsiteisatGrubbo7,shownbythearrow;the entranceislabeled dolinadÂ’ingresso I NVOLVEMENTOFBACTERIAINTHEORIGINOFANEWLYDESCRIBEDSPELEOTHEMINTHEGYPSUMCAVEOF G RAVE G RUBBO (C ROTONE ,I TALY ) 10 N JournalofCaveandKarstStudies, April2012

PAGE 13

tubewassetat50kVand30mA.The2tscanrangewas between22 u and50 u ;andeachscanwasdoneinstepsof 0.05 u .Acountingtimebetweenoneandtensecondsperstep wasselected,dependingonthesampledensity.ThecrystallinephaseswereidentifiedusingtheICDD(International CenterforDiffractionData)databasecards(JCPDS 5 Joint CommiteeonPowderDiffractionStandards). Weusedasimpleprocedureforrecognizingthe crystallinephasespresentineachspecimen.Culturedsolid mediasamplesweredriedat22 u Cor32 u C.Agarmedium wascutinto10 3 30mmflatblocks,0.5mmhigh,and thoserichestincrystalliteswerefixedonadhesivetape. Thisself-supportingfilmwasputintothecenterofthe diffractometerusingaU-shapedsampleholdertominimize backgroundsignals.Forcrystallite-poorsamples,theagar wasdissolvedandcrystalswerecollectedaccordingtothe Rivadeneyraetal.method(1998).Thewashedcrystals wereair-driedat37 u Candthensupportedonaglassslide forX-raymeasurements. Morphologycharacteristicswerestudiedbyscanning electronmicroscopy(SEM).Culturedsolid-mediasamples weredriedat37 u C;agarmediumwascutintoflatblocks. Forcrystallite-poorsamples,theagarwasdissolvedand crystalswerecollectedandpurifiedaccordingtothe Rivadeneyraetal.method(1998).Toobservetheinner portionofthebioliths,crystalswerefirstpowderedusinga mortarandpestle.Allsamplesweregold-shadowedand observedwithaPhilipsSEMXL30CP. For d 13 Cand d 18 Oanalyses,carbonates(1mg)obtained invitroorfromGraveGrubboCaveweredissolvedat70 u CinaKIELIIcarbonatedevice(modifiedbyMcCrea, 1950).TheCO 2 producedwasanalyzedusingaFinniganMat252massspectrometer.DatarelativetotheV-PDB standardarereportedusingtheconventionalnotationas previouslyspecified(Cacchioetal.,2004). Bacterialisolatesdisplayingcalcificationwerealso testedforcalciumcarbonatedissolutionbygrowingcoloniesonDeveze-Brunimedium(pH6.8)containing0.14% or2.5%CaCO 3 at32 u C(NORMALCommision,1990). Wequantifiedcarbonatesolubilizationafter7,15,and 30daysbymeasuringthediameteroftheclearhalothat surroundedeachcolonyinresponsetodecreasedpH (Martinoetal.,1992). R ESULTSAND D ISCUSSION Microbiologicalanalysisofthesamplefromtheupper layerofthecalcareousspeleothemfromGrubboCaveled totheisolationofsixteenbacterialisolates.Eightofthese wereobtainedfromtheoutersub-sampleandeightwere fromtheinnersub-sample.Basedonthecolonymorphologiesoftheisolates,weconcludedthateachsubsamplecontainedthesameeightisolatesandnumbered themfromG1toG8.Theseresultsdemonstratetheexistenceofamicrofloraofculturableheterotrophicbacteria thatwasautochthonoustothissubstrate.ThetwosubsamplescontainedsimilardensitiesofheterotrophicbacteriaculturableintheB-4medium,3.4 3 10 4 cfug 2 1 (dry weigh)fortheoutersampleand2.0 3 10 4 cfug 2 1 (dry weight)fortheinnersample,whichindicatedthatthe isolateswerehomogeneouslydistributedthroughoutthe calcareousbody.Asimilarmicrobialdensityhasbeen reportedforthecalcareousspeleothemsoftheStiffeand CervoCaves(Cacchioetal.,2003band2004,respectively).Theabundanceofculturablebacterialcellsisolated fromtheGraveGrubbospeleothemsuggeststhattheir presencewasnotaccidental. Theisolateswerepresentinsimilarproportionsinthe innerandoutersub-samples.TheG4isolatewasthemost abundant,representing81%ofthepopulation,whereasG5 represented11%.ThecontributionsoftheG1,G2,G3,G6, G7andG8isolatestotheoverallpopulationrangedfrom 0.3to5.4%.Theseresultsareinteresting,butmaynotreflect theactualmicrobialactivitytakingplaceinthespeleothem, ascultivationtechniquesarethoughttogreatlyunderestimatemicrobialdiversityduetothenon-culturabilityofthe largemajorityofmicroorganisms(Dojkaetal.,2000). Theresultsofthecharacterizationtestsandthespecies weidentifiedareshowninTable1.Mostisolateswere aerobicrod-shapedbacteria,ofwhich50%wereGrampositive.Themostabundantisolate,G4,wasanaerobic, spherical-shapedGram-positivebacterium.Ithasbeen reportedintheliteraturethatGram-positivecocciand coccoidformsareextremelycommoninsoilandinundisturbedenvironmentssuchastheGraveGrubboand Cervocaves(Cacchioetal.,2004).Allthecalcifying bacteriareducednitrateandhydrolyzedurea.Bothofthese Table1.IdentificationandculturetraitsofthecalcifyingisolatesfromGraveGrubboCave(Calabria,SouthernItaly). StrainsMicroorganismGramCatalaseOxidaseUreaseNO { 3 ? NO { 2 Anaerobiosis G1 Bacillusmagaterium ++ 2 +++ / 2 G2 Burkholderia sp. 2 + 2 +++ / 2 G3 Pasteurella sp. 2 + 2 ++ 2 G4 Staphylococcus sp. ++ 2 ++ 2 G5 Burkholderia sp. 2 + 2 ++ 2 G6 Brevibacillusbrevis ++ 2 ++ 2 G7Notidentified 2 + 2 +++ / 2 G8 Actinomyces sp. ++ 2 + nd 2 Note: + / 2 5 minimalgrowthandnd 5 notdetermined. P.C ACCHIO ,C.E RCOLE ,R.C ONTENTO ,G.C APPUCCIO ,M.P.M ARTINEZ ,M.D EL .G ALLOAND A.L EPIDI JournalofCaveandKarstStudies, April2012 N 11

PAGE 14

metabolicactivitiesmightbeinvolvedintheprecipitation ofcalciumcarbonateincaves(Castanieretal.,1999). Possiblesourcesofnitratesincavesincludebatguano, organic-richammoniaorammoniumionscarriedinfrom surfacesoils,ammonium-ureafromamberat(caverat urine),bacterialnitrogenfixation,fertilizers,volcanic rocks,andforestlitter(BartonandNorthup,2007). TheAPISystemidentifiedsevenoftheeightcalcifyingbacterialisolatesobtainedfromtheGraveGrubbo speleothemtothegenuslevel(Table1).Thespeciesof50% oftheisolateswerescoredwithlowdiscriminationbecause theAPItestbank,whichisdesignedtoidentifypathogenic species,occasionallymisidentifiessubsurfaceisolates(Amy etal.,1992).IsolatesG1andG6wererelatedto Bacillus genu s( Bacillusmegaterium and Brevibacillusbrevis ,respectively). Bacillus strainswerethemostcommoncalcifying bacteriafoundintheStiffeCave(Cacchioetal.,2003b). IsolatesG2andG5wereidentifiedas Burkholderiasp .The mostabundantisolate,G4,wasrelatedto Staphylococcus sp.IsolateG7wasnotidentified.IsolateG8,whichwasthe onlyonedisplayingpseudomyceliargrowth,wasrelatedto Actinomyces sp. AmatiandGualandi(1934)isolated Bacillusviolaceus and Micrococcusflavusliquiefaciens fromsamplestaken fromtheGortaniCave,whichisinagypsumoutcropnear Bologna.Astudyofthemicrobialpopulationsinthe NovellaCave(Farneto,GessiBolognesiRegionalPark) identifiedmanybacterial( Bacillus sp., Serratia sp.,and Acinetobacter sp.)andmold( Cryptococcus sp., Penicillium sp., Mucor sp.,and Candida sp.)speciesandtheirconnectionstothecave. Underlaboratoryconditions,100percentoftheisolates formedcrystallineCaCO 3 inB-4medium.Thispercentage ishigherthanwhatwefoundinsamplesfromtheStiffeand CervoCaves(96%and75%,respectively)(Cacchioetal., 2003a,b,2004).Thisresultisconsistentwiththoseof severalotherstudiesandconfirmsthatinappropriate conditionsmanybacteriaarecapableofformingCaCO 3 crystals(Boquetetal.,1973). CaCO 3 precipitationoccurredatalltemperatures tested,4,15,22,and32 u C.Theinitiationofprecipitation byallthecalcifyingisolatestooklongerat4,15,and22, thanat32 u C.Allofthecalcifyingisolatesbeganto precipitateCaCO 3 afterthreeweeksat32 u C(Fig.4)and aftersixweeksat4,15,and22 u C(datanotshown).No crystalsweredetectedintheuninoculatedcontrolsorinthe controlsinoculatedwithautoclavedbacterialcells. TherelationshipsbetweenCaCO 3 yieldandtemperatureunderlaboratoryconditionsdependedontheisolate Figure4.RelationshipbetweenthepercentageofstrainsthatwerecalcifyingonB-4solidmediumat4,15,22,and32 6 Cand thenumberofdaysinculture.G7andG8didnotsurviveandarenotincludedinthepercentages. Figure5.InvitroCaCO 3 yieldat4,15,22,and32 6 C.The experimentaldryweightsarethemeansoffourvalues obtainedfromindependentexperiments. I NVOLVEMENTOFBACTERIAINTHEORIGINOFANEWLYDESCRIBEDSPELEOTHEMINTHEGYPSUMCAVEOF G RAVE G RUBBO (C ROTONE ,I TALY ) 12 N JournalofCaveandKarstStudies, April2012

PAGE 15

(Fig.5).Themostabundantisolates,G4andG5,which belongedtothe Staphylococcus and Burkholderia genera, respectively,producedthelargestamountofcrystalsat15 u C(Fig.5).Inapreviousreport,wesuggestedthatthe abilitytoprecipitateCaCO 3 maybeadvantageousand subjecttoevolutionaryselectioninacaveenvironment (Cacchioetal.,2003b).Calciumcarbonatedepositionis usefultobacteriacellsthatremaininsitutakingadvantage ofnutritivemoleculespresentinthewater,likecoralsdo. InthecaseofthespeleothemoftheGraveGrubboCave, carbonatogenesismightkeepthecalcifyingbacteriain place,andcalcifyingcellsmayavoidbeingwashedoutby theflowoftheundergroundriver. ThecalciumcarbonateyieldsofthestrainsG7andG8 arenotavailable.Theseisolatessufferedduringthetransitionfromthestarved,oligotrophicconditionsofcave environmentstotheeutrophicconditionsofnutrientagar anddiedbeforecrystalscouldberecovered.Itmaybe difficultformicroorganismsfromnutrient-poorcaveenvironmentstoadapttothesuddenpresenceofnutrientsin alaboratoryenvironment,andtheymaysimplydiefrom osmoticstress(Koch,1997).Inadaptingtocaveenvironments,microorganismsemployelaboratemechanismsto pullscarcenutrientsintothecell.Whenthesehighlyadaptedorganismsarethenexposedtotherichnutrientsof laboratorymedia,theseextremescavengingmechanisms Figure6.Carbonisotopecomposition( % )versustemperature( 6 C)ofcalciumcarbonatesproducedinvitrobythecalcifying bacterialisolates. Figure7.Oxygenisotopiccomposition( % )versustemperature( 6 C)ofcalciumcarbonatesproducedinvitrobythecalcifying bacterialisolates. P.C ACCHIO ,C.E RCOLE ,R.C ONTENTO ,G.C APPUCCIO ,M.P.M ARTINEZ ,M.D EL .G ALLOAND A.L EPIDI JournalofCaveandKarstStudies, April2012 N 13

PAGE 16

continueunabatedandthebacteriaquicklygorgethemselvestodeath(Koch,1997,2001). Xray-diffractionanalysisshowedthatallthecalcifying bacteriaproducedcalciteandthat Burkholderia sp.also depositedvateriteintraceamounts( 1%)at15 u C(data notshown).Vateriteprecipitationwasbothspecies-specific andtemperature-dependent.XRDanalysisofthespeleothemrevealedthepresenceofcalciteinlargeamounts, quartzinsmallamounts,andaragoniteandgibbsitein traces.Noneoftheisolatesprecipitatedcalciumcarbonate intheformofaragonite. Theisotopicanalysisperformedonthecarbonate concretionofthespeleothemhasshownmediumvalues for d 13 Cand d 18 Oof 2 6.4 % and 2 5.7 % ,respectively. Previousstudies(Turi,1986)haveshownthatvaluesfor d 13 Clowerthan 2 4 % indicatethatinadditionto‘‘heavy’’ carbonfromthedissolutionofcarbonaticrocks,‘‘light’’ biogeniccarbon,whichisplant-derivedCO 2 dissolvedin thesoil,mustalsocontributetocalciteprecipitation.On theotherhand,manyofthecarbonaticconcretionsof microbialoriginshow d 13 Cvaluessimilartothevalues foundintheGraveGrubbospeleothem(Andrewsetal., 1997).The d 18 Ovaluesarecompatiblewithprecipitation atequilibriumfrommeteoricwateratcavetemperature (O’Neiletal.,1969).Fromtheseobservations,oxygenand carbonfractionationduetothepresenceofcalcifying bacteriacannotberuledout,althoughtheenvironmental conditions(temperature, d 18 Ointhewater,carbonderived frominorganicandorganicsources,etc.)areessentialin definingtheisotopiccompositionofthecalciteconcretions ofcaves.Inaddition,analysisofthecarbonandoxygen isotopesofcarbonatesproducedinvitroshowedthat,first, theisotopiccompositionofthecalciumcarbonatesproducedbythedifferentbacterialisolatesdependedmoreon temperaturethanonthetypeofbacterialisolate(exceptfor G3)and,second,atcavetemperature(15 u C),themost abundantisolate, Staphylococcus sp.,producedthelargest amountofcrystalsandprecipitatedthecarbonatesmost depletedin 13 Cand 18 O(Figs.6and7).Apositivecorrelationexistedbetweenthe d 13 Cand d 18 Ovaluesofthe crystals(Fig.8). Itwasnotpossibletocomparethe d 13 Cand d 18 Ovalues obtainedofthecarbonatesproducedinvitrowiththevalues oftheconcretionsofthecavebecausethecompositionofthe cave’ssourcesofcarbonandoxygenareunknown. Bacillusmegaterium isolatesprecipitatedvariousbiomineralsinvitro,predominantlyspherulitesconsistingof isolatedspheresorforminggroupsofspheres(Fig.9a,b,c). Biofilmswereobservedasnon-calcareousbridgesamong thecrystals(Fig.9b,c).Hemisphericalformswerealso presentwithradialinnersurfaces(Fig.9d,e,f).Theouter surfacesofthesestructureswerefrequentlyroughdueto thedenovodepositionofcalciumcarbonate(Fig.9a,d,e). Inapreviousstudy,weobtainednineisolatesof Bacillus megaterium (Cacchioetal.,2003b).SEManalysisrevealed thatsomeofthoseisolatesproducedsphericalorhemisphericalcrystalssimilartothoseproducedbytheG1 isolatedhere.IsolateG6, Brevibacillussp. ,alsoproduced sphericalcrystals. Pasteurella sp.precipitatedirregularly shapedbiolithsthattendedtoaggregate. Staphylococcus sp.depositedglobularcrystalswitharoughoutersurface. Burkholderia isolatesproducedellipticalcrystalswitha scalysurface.Theinnerportionsofthesebiolithscontained imprintsthatweresimilarinsizeandshapetothebacterialcells(Fig.9g,h,i,l).Wedidnotanalyzethecarbonates producedbytheG7isolateand Actinomyces sp.bySEM. Thedifferentmorphologiesoftheprecipitatesformedby thedifferentcalcifyingisolatesconfirmedthatcrystalmorFigure8.Correlationbetween d 13 Cand d 18 Ovalues( % )of thecalciumcarbonatesprecipitatedinvitrobythecalcifying isolates. R Figure9.ScanningelectronmicrographsofthebiolithsprecipitatedonB-4agarby Bacillusmegaterium (a,b,c,d,e,f)and Burkholderia sp.(g,h,i,l)isolatedfromthenovelcalcitestructureinGraveGrubboCave,SouthernItaly. a )Twoindividual intergrowingbioliths;scalebar20 m m. b) Alargeaggregateofindividualcrystalsboundbymeansofnon-globularcarbonate bridges;scalebar20 m m. c) Ahighermagnificationofthebiofilm;scalebar10 m m. d) Radiallyarrangedhemispherical calciumcarbonatecrystal;scalebar20 m m. e) Theoutersurfaceofthebiolithcharacterizedbydenovodepositionofcalcium carbonate;scalebar20 m m. f) Fragmentofacrystalshowingathree-layeredstructure;scalebar20 m m. g) Theinnersurface I NVOLVEMENTOFBACTERIAINTHEORIGINOFANEWLYDESCRIBEDSPELEOTHEMINTHEGYPSUMCAVEOF G RAVE G RUBBO (C ROTONE ,I TALY ) 14 N JournalofCaveandKarstStudies, April2012

PAGE 17

showsmicrobialimprintssimilarinsizeandshapetothemicrobialrods;scalebar20 m m. h) Ahighermagnificationofthe internalporousstructureofthebiolith;scalebar20 m m. i) Calciumcarbonatelayerwithbacterialimprintsarrangedina geometricstreamdisposition;scalebar50 m m. l) Detailof i; scalebar20 m m. P.C ACCHIO ,C.E RCOLE ,R.C ONTENTO ,G.C APPUCCIO ,M.P.M ARTINEZ ,M.D EL .G ALLOAND A.L EPIDI JournalofCaveandKarstStudies, April2012 N 15

PAGE 18

phologywasspecies-specific,suggestingthatthebacteria playamajorroleintheprecipitationprocess. Somecrystalsexhibitedbacterialimprintsarrangedina geometricstreamdisposition(Fig.9g,i,l).Thisparticular imprintpatternmaybeduetotheinvolvementofquorum sensingincolonygrowth(Brandaetal.,2004).Quorum sensingisaprocessofcell-to-cellcommunicationthatbacteriausetoassesstheirpopulationdensitytocoordinategene expressionofthecommunity(MillerandBassler,2001). Quorumsensingrequiresproduction,secretion,anddetectionofextracellularsignalmoleculescalledautoinducers, someofwhichareusedforintraspeciescommunication, whileotherspromoteinterspeciescommunication(Federle andBassler,2003).Manygram-negativebacterialspeciesuse acylhomoserinelactones(AHLs)asautoinducers.Quorum sensingcontrolsseveralimportantfunctionsinbacteria, includingtheproductionofvirulencefactorsandbiofilm formation.Wehypothesizethatthiscell-to-cellcommunicationmaybealsoinvolvedinthecalcificationprocess.Tothis aim,weproducedanAHL 2 mutantofthestrainPHP7of Burkholderiacepacia. TheAHL 2 mutantgrownonB-4agar plateslostthecalcifyingcapabilityofthewildtypestrain (preliminarydatanotpublished). TheSEManalysisofthespeleothemmaterialconfirm thelayeredorganizationofthetubebodythatwasapparentuponobservation(Fig.3),supportingthetubular structureofthecalcareousbodyandthespongystructure ofthethree-layeredwall(Fig.10a,b).Calcifiedfilamentous bacterialcellsweredetectedbetweentheouterwallofthe tubesandthedepositedcalcite(Fig.10c),whichpartially occludedthelumenofthetubes(Fig.10a).Finally,only theG7isolatewasabletosolubilizeCaCO 3 (0.14%),after oneweekat32 u C. C ONCLUSIONS Withinthepasttwodecades,growinginterestincave microbiotahashelpedtorecognizeandunderstandthe importanceofmicrobialspeciesincaves.Studiesofcave microbiologyhaveidentifiedawidevarietyofdifferent cave-dwellingmicroorganisms,revealedhowtheseorganismsinteractwithandadapttothecaveenvironment,and disclosedtheirrolesincreatinganddestroyingsecondary mineraldeposits.Mostoftheresearchinvolvingthe gypsumareasinItalyhasbeencarriedoutsince2000on samplesfromcaveslocatedinthelargegypsumareasof Sicily,Calabria,andEmilia. Microbescyclenutrientswithincaveenvironments.The processesofammonification,nitrification,denitrification, andnitrogenfixationhaveallbeendocumentedtooccurin caves.Sulfideandsulfuroxidizersandsulfatereducersare alsofoundincaves.Nitrogenandthesulfurcyclesmay bothoccur,separatelyorconcurrently,duringheterotrophicbacterialcarbonatogenesisincaves. Microbesarecriticaltofood-limitedcaveenvironmentsbecausetheymayactasprimaryproducersinan environmentlackingphotosyntheticorganisms. Studiesofsulfurbacteriahaverapidlyexpandedinthe lastdecade,reflectingtherecentdiscoveriesofsulfurcontainingcaves,includingMovileCaveinRomania.The Figure10.SEMimagesofcavecrystalscollectedfromthe speleotheminGraveGrubboCave. a) Cross-sectionofthe aggregateshowingtheouter-innerlaminaeandthemiddle, sponge-likesepta. b) Detailof a c ,Spikecrystalsdeposited aftertubeformationarelikelytooccludethelumenofthe tube.Filamentouscalcifiedbacteriainteractingwiththe calcitelayer(arrow). I NVOLVEMENTOFBACTERIAINTHEORIGINOFANEWLYDESCRIBEDSPELEOTHEMINTHEGYPSUMCAVEOF G RAVE G RUBBO (C ROTONE ,I TALY ) 16 N JournalofCaveandKarstStudies, April2012

PAGE 19

foodwebofthiscaveisbasedonchemolithotrophicprocessesinvolvingsulfurandothercompounds.Amongthe Italiancavesinthegypsumarea,aremarkablenumberof sulfurbacteriahavebeenisolatedinthecavesofSanta Ninfa,Sicily.Sulfurbacteriahavealsobeendetectedin GraveGrubboCave,whereasubterraneanecosystemdriven bythesechemoautotrophicmicroorganisms,astrophic symbiontsoftroglobiticinsectsabletoagglutinatelithoid elements,hasbeenhypothesizedtobeindirectlyinvolvedin thedepositionoftheunusualspeleothemfoundinthiscave. Westudiedthispeculiarlystructuredcavespeleothem todeterminewhetherheterotrophicbacterialcarbonatogenesiswasinvolvedinitsformation.TheGraveGrubbo speleothemwasfoundtocontainalargenumberof culturableheterotrophicmicroorganisms,whosepresence waslikelynotaccidental.Alltheisolatesobtainedfromthe speleothemsamplesprecipitatedCaCO 3 invitroinagar containingcalciumacetate.Therelativeabundanceofeach isolatewasfoundtoberelatedtocarbonateyieldat15 u C, suggestingthatcalcificationmaybeaselectiveadvantage. Isotopicanalysisshowedthat,aspreviouslysuspected, the d 13 CvaluesoftheGraveGrubbospeleothemwere compatiblewithabiogenicoriginandthattheCaCO 3 biolithsproducedinvitrobythe Staphylococcus sp.,the mostabundantisolateandtheonethatproducethelargest amountsofcarbonatesinvitroat15 u C,had d 13 Cand d 18 O valuesthatwerelighterthanthecarbonatesproducedin vitrobytheothercalcifyingisolates.Inconclusion,these isotopicdatawereconsistentwithcalcifyingbacteriamaking asignificantcontributiontothedepositionofspeleothem carbonates. SEMimagesrevealedthepresenceofcalcifiedbacterial cellsinteractingwiththecalcitesubstrateofthespeleothem,thepresenceofabiofilmamongthebioliths obtainedinvitro,thepresenceofbacterialimprintsinthe interioroftheCaCO 3 biolithsproducedinvitro,and differentspecies-specificmorphologiesamongthebioliths obtainedinvitro.Atypicalstream-likeorganizationofthe bacterialimprints,observedinsomeinstances,suggests thataquorum-sensingsystemmaybeinvolvedinthe calciumcarbonateprecipitationprocess. Ourdataclearlydemonstratethatcalcifyingbacteria isolatedfromthisunusualcalcareousspeleothemhostedin theGraveGrubboCaveareabletoprecipitateCaCO 3 in vitroandmayplayanimportantroleinthedepositionof thesecavecrystals. A CKNOWLEDGEMENTS WethankG.FerriniandA.Morettiforprovidingsamples fromtheGraveGrubboCaveandfortheassistantwiththe geologicalconsiderations,andC.VeroliandM.Molafor theircollaborationinpowderdiffractionmeasurementsand stableisotopeanalysis,respectively.WealsothankM. GiammatteoandL.ArrizzaforassistancewithSEMandR. DiStefanoforassistancewiththeCalabriamap. R EFERENCES Amati,A.,andGualandi,C.,1934,Lamicrofloradialcuneacque cavernicoledelsottosuolobolognese,Naples,Riv.diFisica, MatematicaeScienzeNaturali,ser.3,16p. Amy,P.S.,Haldeman,D.L.,Ringelberg,D.,Hall,D.H.,andRussel,C., 1992,Comparisonofidentificationsystemsforclassificationof bacteriaisolatedfromwaterandendolithichabitatswithinthedeep subsurface:AppliedandEnvironmentalMicrobiology,v.58,no.10, p.3367–3373. Andrews,J.E.,Riding,R.,andDennis,P.F.,1997,Thestableisotope recordofenvironmentalandclimaticsignalsinmodernterrestrial microbialcarbonatesfromEurope:Palaeogeography,Palaeoclimatology,Palaeoecology,v.129,p.171–189,doi:10.1016/S0031-0182 (96)00120-4. Barton,H.A.,andNorthup,D.E.,2007,Geomicrobiologyincave environments:past,currentandfutureperspectives:JournalofCave andKarstStudies,v.69,no.1,p.163–178. Baskar,S.,Baskar,R.,Lee,N.,andTheophilus,P.K.,2009,Speleothems fromMawsmaiandKremPhyllutcaves,Meghalaya,India:some evidencesonbiogenicactivities:EnvironmentalGeology,v.57,no.5, p.1169–1186,doi:10.1007/s00254-008-1413-y. Baskar,S.,Baskar,R.,Mauclaire,L.,andMcKenzie,J.A.,2006, Microbiallyinducedcalciteprecipitationincultureexperiments: PossibleoriginforstalactitesinSahastradharacaves,Dehradun, India:CurrentScience,v.90,p.58–64. Boquet,E.,Boronat,A.,andRamos-Cormenzana,A.,1973,Production ofcalcite(calciumcarbonate)crystalsbysoilbacteriaisageneral phenomenon:Nature,v.246,p.527–529,doi:10.1038/246527a0. Branda,S.S.,Gonza les-Pastor,J.E.,Dervyn,E.,Ehrlich,S.D.,Losick,R., andKolter,R.,2004,Genesinvolvedinformationofstructured multicellularcommunitiesby Bacillussubtilis :JournalofBacteriology, v.186,p.3970–3979,doi:10.1128/JB.186.12.3970-3979.2004. Cacchio,P.,Contento,R.,Ercole,C.,Cappuccio,G.,PreiteMartinez,M., andLepidi,A.,2004,Involvementofmicroorganismsintheformationof carbonatespeleothemsintheCervoCave(L’Aquila-Italy):GeomicrobiologyJournal,v.21,p.497–509,doi:10.1080/01490450490888109. Cacchio,P.,Ercole,C.,andLepidi,A.,2003a,Calciumcarbonatedeposition inlimestonecaves:microbiologicalaspects:SubterraneanBiology,v.1, p.57–63. Cacchio,P.,Ercole,C.,Cappuccio,G.,andLepidi,A.,2003b,Calcium carbonateprecipitationbybacterialstrainsisolatedfromalimestone caveandfromaloamysoil:GeomicrobiologyJournal,v.20, p.85–98,doi:10.1080/01490450303883. Can averas,J.C.,Hoyos,M.,Sanchez-Moral,S.,Sanz-Rubio,E.,Bedoya, J.,Soler,V.,Groth,I.,Schumann,P.,Laiz,L.,Gonzales,I.,andSaizJimenez,C.,1999,Microbialcommunitiesassociatedwithhydromagnesiteandneedle-fiberaragonitedepositsinakarsticcave (Altamira,northernSpain):GeomicrobiologyJournal,v.16, p.9–25,doi:10.1080/014904599270712. Can averas,J.C.,Sanchez-Moral,S.,Soler,V.,andSaiz-Jimenez,C.,2001, Microorganismsandmicrobiallyinducedfabricsincavewalls: GeomicrobiologyJournal,v.18,p.223–240,doi:10.1080/01490450 152467769. Castanier,S.,LeMetayer-Levrel,G.,andPerthuisot,J.-P.,1999,Cacarbonatesprecipitationandlimestonegenesis—themicrobiogeologistpointofview:SedimentaryGeology,v.126,p.9–23,doi:10.1016/ S0037-0738(99)00028-7. Castanier,S.,LeMe tayer-Levrel,G.,andPerthuisot,J.-P.,2000,Bacterial rolesintheprecipitationofcarbonateminerals, in Riding,R.E.,and Awramik,S.M.,eds.,MicrobialSediments,Berlin,Springer-Verlag, p.32–39. Dojka,M.A.,Harris,J.K.,andPace,N.R.,2000,Expandingtheknown diversityandenvironmentaldistributionofanunculturedphylogeneticdivisionofbacteria:AppliedandEnvironmentalMicrobiology, v.66,no.4,p.1617–1621. Ehrlich,H.L.,2002,Geomicrobiology,4thedition,CRCPress,768p. Ercole,C.,Cacchio,P.,Botta,A.L.,Centi,V.,andLepidi,A.,2007, Bacteriallyinducedmineralizationofcalciumcarbonate:theroleof exopolysaccharidesandcapsularpolysaccharides:Microscopyand Microanalysis,v.13,p.42–50,doi:10.1017/S1431927607070122. Federle,M.J.,andBassler,B.L.,2003,Interspeciescommunicationin bacteria:TheJournalofClinicalInvestigation,v.112,p.1291–1299, doi:10.1172/JCI20195. P.C ACCHIO ,C.E RCOLE ,R.C ONTENTO ,G.C APPUCCIO ,M.P.M ARTINEZ ,M.D EL .G ALLOAND A.L EPIDI JournalofCaveandKarstStudies, April2012 N 17

PAGE 20

Ferrini,G.,ed.,1998,L’areacarsicadellevigne(Verzino-Crotone)Studio multidisciplinare(Thekarstareaof‘‘LeVigne’’–VerzinoCrotonemultidisciplinaryresearches):MemorieIstitutoItalianodiSpeleologia. Ferrini,G.,1998,LitostratigrafiadelcomplessocarsicoGraveGrubboRisorgivaValloneCufalo(GessoarenitiMessinianedelBacino Crotonese):Memoriedell’IstitutoItalianodiSpeleologia,ser.2, v.10,p.47–52. Forti,P.,andLombardo,N.,1998,Idepositichimicidelsistemacarsico GraveGrubbo–RisorgentediValloneCufalo(Verzino,Calabria): Memoriedell’IstitutoItalianodiSpeleologia,v.10,p.83–92. Forti,P.,Galdenzi,S.,andSarbu,S.M.,2002,Thehypogeniccaves:a powerfultoolforthestudyofseepsandtheirenvironmentaleffects: ContinentalShelfResearch,v.22,no.16,p.2373–2386,doi:10.1016/ S0278-4343(02)00062-6. Galdenzi,S.,andMenichetti,M.,1998,Aspettimorfologiciedevolutividi unsistemacarsicoipogeoneigessidiVerzino(Calabria):Memorie dell’IstitutoItalianodiSpeleologia,ser.2,v.10,p.71–82. Groth,I.,Schumann,P.,Laiz,L.,Sanchez-Moral,S.,Can averas,C.J., andSaiz-Jimenez,C.,2001,GeomicrobiologicalstudyoftheGrotta deiCervi,PortoBadisco,Italy:GeomicrobiologicalJournal,v.18, p.241–258,doi:10.1080/01490450152467778. Groth,I.,Vettermann,R.,Schuetze,B.,Schumann,P.,andSaiz-Jimenez, C.,1999,ActinomycetesinkarsticcavesofnorthernSpain(Altamira andTitoBustillo):JournalofMicrobiologicalMethods,v.36, p.115–122,doi:10.1016/S0167-7012(99)00016-0. Koch,A.L.,1997,Microbialphysiologyandecologyofslowgrowth: MicrobiologyandMolecularBiologyReviews,v.61,p.305–318. Koch,A.L.,2001,Oligotrophsversuscopiotrophs:BioEssays,v.23, p.657–661,doi:10.1002/bies.1091. Laiz,L.,Gonzalez-Delvalle,M.,Hermosin,B.,Ortiz-Martinez,A.,and Saiz-Jimenez,C.,2003,Isolationofcavebacteriaandsubstrate utilizationatdifferenttemperatures:GeomicrobiologyJournal,v.20, p.479–489,doi:10.1080/713851125. Laiz,L.,Groth,I.,Gonzales,I.,andSaiz-Jimenez,C.,1999,MicrobiologicalstudyofthedrippingwatersinAltamiracave(Santillanadel Mar,Spain):JournalofMicrobiologicalMethods,v.36,p.129–139, doi:10.1016/S0167-7012(99)00018-4. Laiz,L.,Groth,I.,Schumann,P.,Zezza,F.,Felske,A.,Hermosin,B.,and Saiz-Jimenez,C.,2000,MicrobiologyofthestalactitesfromGrottadei Cervi,PortoBadisco,Italy:InternationalMicrobiology,v.3,p.25–30. LeMe tayer-Levrel,G.,Castanier,S.,Loubie `re,J.F.,andPerthuisot,J.P., 1997,Lacarbonatogene `sebacte riennedanslesgrottes.E tudeauMEB d’unehe lictitedeClamouse,He rault,France:ComptesRendusde l’Acade miedesSciencesSerieIIa:SciencesdelaTerreetdesPlanets, v.325,p.179–184,doi:10.1016/S1251-8050(97)88286-9. McCrea,J.M.,1950,Ontheisotopicchemistryofcarbonatesanda palaeotemperaturescale:JournalofChemicalPhyics,v.18,p.849– 857,doi:10.1063/1.1747785. Martino,T.,Salamone,P.,Zagari,M.,andUrz `,C.,1992,Adesionea substratisolidiesolubilizzazionedelCaCO 3 qualemisuradella capacita `deteriorantedibatteriisolatidalmarmoPentelico, in Italian SocietyofGeneralMicrobiologyandMicrobialBiotechnology (SIMGBM)XIMeeting,Gubbio;p.249–250. Melim,L.A.,Shinglman,K.M.,Boston,P.J.,Northup,D.E.,Spilde, M.N.,andQueen,J.M.,2001,Evidenceformicrobialinvolvementin poolfingerprecipitation,HiddenCave,NewMexico:GeomicrobiologyJournal,v.18,p.311–329,doi:10.1080/01490450152467813. Miller,M.B.,andBassler,B.L.,2001,Quorumsensinginbacteria:Annual ReviewofMicrobiology,v.55,p.165–199,doi:10.1146/annurev.micro.55.1.165. NORMALCommission,1990,NORMAL9/88recommendations.MicrofloraAutotrofaedEterotrofa:TecnichediIsolamentoinColtura: Rome,ConsiglioNazionaleRicerche,IstitutoCentraleRestauro,26p. Northup,D.E.,Dahm,C.N.,Melim,L.A.,Spilde,M.N.,Crossey,L.J., Lavoie,K.H.,Mallory,L.M.,Boston,P.J.,Cunningham,K.I.,and Barns,S.M.,2000,Evidenceforgeomicrobiologicalinteractionsin Guadalupecaves:JournalofCaveandKarstStudies,v.62,p.80–90. O’Neil,J.R.,Clayton,R.N.,andMayeda,T.K.,1969,Oxygenisotope fractionationindivalentmetalcarbonates:JournalofChemical Physics,v.51,p.5547–5558,doi:10.1063/1.1671982. Poluzzi,A.,andMinguzzi,V.,1998,Uncasodibiocostruzioneinun ambientedigrotta.L’areacarsicadelleVignediVerzino:Memorie dell’IstitutoItalianodiSpeleologia,ser.2,v.10,p.93–100. Riding,R.,2000,Microbialcarbonates:thegeologicalrecordofcalcified bacterial-algalmatsandbiofilms:Sedimentology,v.47,no.suppl.S1, p.179–214,doi:10.1046/j.1365-3091.2000.00003.x. Rivadeneyra,M.A.,Delgado,G.,Ramos-Cormenzana,A.,andDelgado, R.,1998,BiomineralisationofcarbonatesbyHalomonaseurihalina insolidandliquidmediawithdifferentsalinities:crystalformation sequence:ResearchinMicrobiology,v.149,p.277–287. Simkiss,K.,andWilbur,K.M.,1989,Biomineralization:CellBiologyand MineralDeposition,SanDiego,AcademicPress,337p. Turi,B.,1986,Stableisotopegeochemistryintravertines, in Fritz,P.,and Frontes,J.C.,eds.,HandbookofEnvironmentalIsotopeGeochemistry 2:TheTerrestrialEnvironment:Elsevier,Amsterdam,p.207–235. I NVOLVEMENTOFBACTERIAINTHEORIGINOFANEWLYDESCRIBEDSPELEOTHEMINTHEGYPSUMCAVEOF G RAVE G RUBBO (C ROTONE ,I TALY ) 18 N JournalofCaveandKarstStudies, April2012

PAGE 21

THEPREHISTORICCAVEARTANDARCHAEOLOGYOF DUNBARCAVE,MONTGOMERYCOUNTY,TENNESSEE J AN F.S IMEK 1 ,S ARAH A.B LANKENSHIP 1 ,A LAN C RESSLER 2 ,J OSEPH C.D OUGLAS 3 ,A MY W ALLACE 4 D ANIEL W EINAND 1 AND H EATHER W ELBORN 1 Abstract: DunbarCaveinMontgomeryCounty,Tennesseehasbeenusedbypeoplein agreatvarietyofways.Thispaperreportsonprehistoricusesofthecave,whichwere quitevaried.Thevestibuleofthecave,whichistodayprotectedbyaconcreteslab installedduringthecave’sdaysasanhistorictouristshowplace,sawextensiveandvery longtermoccupation.DiagnosticartifactsspantheperiodfromLatePaleo-Indian (ca.10,000-yearsago)totheMississippian,andincludeArchaic(10,000to3,000-years ago)andWoodland(3,000–1,000-yearsago)culturalmaterials.Theseincludea paleoindianBeaverLakePoint,Kirkclusterpoints,LittleRivertypes,Ledbettertypes, numerousstraight-stemmedpointtypes,HamiltonandMadisonprojectilepoints. Woodlandperiodceramicscomprisevariouslimestonetemperedforms,allinlow quantities,andcord-markedlimestonetemperedwaresintheuppermostWoodland layers.Shell-temperedceramicsbearwitnesstoarichMississippianpresenceatthetopof thedeposit.Giventhischronologicalspan,theDunbarCavesequenceisascompleteas anyineasternNorthAmerica.However,problemswithpreviousexcavationstrategies makemuchoftheexistingarchaeologicalrecorddifficulttointerpret.Wepresentanew seriesofradiocarbonagedeterminationsthatshowboththegreattimedepthofthe vestibuledepositsandtheproblemswiththeirintegrity.Therewasalsoextensive prehistoricuseofDunbarCave’sdarkzone,includingmineralextraction,andritual intermentofthedead.Mostimportantly,thirty-fivepetroglyphsandpictographswere madeonthecavewalls,mostprobablyduringtheMississippianperiod.Theseinclude geometricshapes,abstractcompositions,andhumanfiguresincludingamythological herowarriorknownfromotherexamplesofMississippianiconography.Dunbarmay alsohaveseenritualvisitationveryearly,i.e.,duringtheArchaicperiod(ca.5,000-years ago),entailingtheplacementofofferingsinthecave’sinteriorwaters. I NTRODUCTION Overthepasttwentyyears,archaeologistsfromthe UniversityofTennesseeandelsewhere,alongwithdevoted avocationalcavers,haveworkedtoexpandourknowledge ofprehistoriccaveartintheAppalachianPlateauuplands ofsoutheasternNorthAmerica(SimekandCressler,2005). Firstidentifiedbyarchaeologistsin1980(Faulkneretal., 1984),thiscaveartrepresentsawidespread,complex,and longstandingaspectofindigenousprehistoricculture,one withlocaloriginsanddevelopmentandoneintrinsically linkedtotheevolutionofprehistoricsoutheasternreligious iconography(FaulknerandSimek,1996).Althoughwe havenowexaminedmorethanonethousandsoutheastern cavesinhopesoffindingprehistoricart,therearemore thanninethousandcavesinTennesseealone,withthousandsmoreinAlabama,Georgia,andtheuppersouth (Simek,2010).Thereisagreatdealofsurveyworkstill beforeus,andtherewillcertainlybemorecave-artsites discoveredinthefuture.Inthispaperwewilldiscussthe fiftiethcaveinthecatalog,DunbarCave(40MT43)in MontgomeryCounty,Tennessee(Fig.1). Asfaraswecandetermine,dark-zonecaveart(thatis, decorationintheareasofcavesbeyondthereachof externallight)wasactuallyknownamongasmallgroup ofcaversintheSoutheastfromthe1950’s.Engravings thatwerethoughtbythosecaverswhosawthemtobe prehistoricwereidentifiedatthemouthof12 th Unnamed CaveinTennessee.ThesiteremainedunknowntoarchaeologistsuntilCharlesFaulkneroftheUniversityof Tennesseewastakenthereinthe1980sinconjunction withhisworkatMudGlyphCave(Faulkner,1988).Mud GlyphCaveitselfwasdiscoveredin1979whenarecreationalcaverexploredanarrowsubterraneanstream passageandsawimagesincisedintothewetclayliningthe streambanks(Faulkneretal.,1984;Faulkner,1986).The caveralertedanarchaeologistfriendwhotoldFaulkner abouttheseimages.Uponseeingthesite,Faulknerquickly recognizedthattheartwasprehistoric,andin1980,he initiatedadocumentationproject.MudGlyphCaveart wasseenasresemblingthatfoundonMississippianperiod ceremonialobjects,andtherefore,linkedtothewider *CorrespondingAuthor:jsimek@utk.edu 1 DepartmentofAnthropology,UniversityofTennessee,Knoxville,TN37996-0720 2 UnitedStatesGeologicalSurvey,Atlanta 3 VolunteerStateCommunityCollege,Gallatin,TN37066 4 DunbarCaveStateNaturalArea,Clarksville,TN37043 J.F.Simek,S.A.Blankenship,A.Cressler,J.C.Douglas,A.Wallace,D.Weinand,andH.Welborn–Theprehistoriccaveartand archaeologyofDunbarCave,MontgomeryCounty,Tennessee. JournalofCaveandKarstStudies, v.74,no.1,p.19–32.DOI:10.4311/ 2011AN0219 JournalofCaveandKarstStudies, April2012 N 19

PAGE 22

Mississippianiconography,datingtoafterAD1000and labeledastheSoutheastCeremonialComplex(SECC) (WaringandHolder,1945).Othersitesquicklybeganto cometolight.By1988,Faulknerdocumentedsevencaveartsites,includingMudGlyphand12 th Unnamedcaves (Faulkner,1988). SincethediscoveryofMudGlyphCave,darkzoneart hasbeenrecordedinsixty-nineothercavesinTennessee, Alabama,Florida,Georgia,Kentucky,Virginia,andWest Virginia.Therearealsoafewcave-artsitesinthe MississippiValley,includingArkansas(Sabo,2008),Missouri(Diaz-GranadosandDuncan,2000;Diaz-Granados etal.,2001),Illinois(Wagner,1996),andWisconsin (Boszhardt,2003).Chronologicaldatafromthesesites demonstratealong-termregionaltraditionofcaveart beginningsome6000yearsago(Creswell,2007;Simekand Cressler,2009).AsFaulknerobserved,someoftheimagery canbeunderstoodintermsofotherprehistoriciconography,i.e.,SECC,butsomehaslessobviousconnectionwith thatimagery.Asweshow,thecaveartinDunbarCaveis markedlySECCinitscontent. T HE P REHISTORIC C AVE A RTOF D UNBAR C AVE InJanuary2004acavinggroupincludingJoseph Douglas(ProfessorofHistoryatVolunteerCommunity CollegeinGallatin,Tennessee,andAdjunctintheDepartmentofAnthropologyattheUniversityofTennessee) andcaveauthorLarryMatthewsvisitedDunbarCave StateNaturalAreainthecompanyofAmyWallace, InterpretiveSpecialistatthepark.Inanareaofthecave knownastheBallroomseveralhundredmetersintothe darkzone,Douglasnoticedtwocharcoaldrawingsonthe walloverlaidbynineteenth-centurygraffiti.(Placesnamed inthispapercanbelocatedbyreferencetoWalter ScheffrahnÂ’s1978mapofDunbarCavebasedonaNorth IndianaGrottosurveyaspublishedinMatthews(2005, p.27).TheBallroom,notlabeledonthatmap,isthe chamberjustnortheastoftheCounterfeitersRoominthe lowercenter).Douglasphotographedwhathesawandsent thephotostoSimekattheUniversityofTennessee. InFebruaryof2005,SimekjoinedDouglas,John FroeschaueroftheTennesseeDepartmentofEnvironment andConservation,AmyWallace,andAlanCresslerofthe U.S.GeologicalSurveyatthesite.Onthatvisit,thetwo glyphsthathadbeendiscoveredinJanuarywereobserved anddocumented.Inaddition,asignificantnumberofnew pictographsandpetroglyphswereidentifiedinthesame areaofthecaveandotherpassages.Wealsoexaminedthe cavefloorforevidenceofprehistoricoccupationandfound humanremainsinasmallalcovenearthecavemouththat hadbeennotedasalootedburialareainearliertesting workbytheTennesseeDivisionofArchaeology(Butler, 1977).Wesawsignificantinsitudepositsofburntriver cane( Arundinaria sp.)thatwereclearlyancienttorch remnants,andstokemarksonthecavewallsreflecting maintenanceofburningtorches. Documentationhasproceededsince2005.Thisinvolves detailedphotographicrecordingoftheglyphsandmapping oftheglyphdistributionsandthegalleriesinwhichthe glyphsarefoundusingatotal-stationlasertransit.The surveycontinuesinotherpartsofthecave,whichismore than8mileslong.Cane-torchstokemarkshavebeen recordedatmanyplacesonthecavewalls,bothnearand awayfromtheglyphpanels.Inplaceswherethesediment floorwasnotdisturbedbyhistoricdigging,torchremnants Figure1.MapshowingthelocationofDunbarCave,Tennessee. T HEPREHISTORICCAVEARTANDARCHAEOLOGYOF D UNBAR C AVE ,M ONTGOMERY C OUNTY ,T ENNESSEE 20 N JournalofCaveandKarstStudies, April2012

PAGE 23

canbefound(Fig.2,Table1).Severaloftheseweretaken as 14 Csamplesundertermsofourstatearchaeological researchpermit. Atotalofthirty-fiveindividualglyphshavebeen recordedsofar,nearlyallconcentratedinfourpanels alongawest-sidewallsegmentintheBallroom.Twoare pictographsofunidentifiableshapesincorporatingdisks. Threeglyphs,twointhemainroomandoneontheceiling attheentrancetothisroom,aresinglecircles.Ineachcase, theseareveryfaintimages,andtheymayactuallyhave Figure2.Burntrivercane( Arundinaria sp.)fragmentsonthesedimentfloorinsideDunbarCave. Table1.RadiocarbonagedeterminationsfromDunbarCave(40MT43).AllassaysperformedbyBetaAnalyticonburntriver cane( Arundinaria sp.)byAtomicMassSpectrometryandcalibratedat2 s withINTCAL04(Reimeretal.,2004).Location descriptionsinthetablefollowWalterScheffrahnÂ’s1978mapofDunbarCavebasedonaNorthIndianaGrottosurvey (Matthews,2005,p.27). Sample Reference Measured AgeBP 13 C/ 12 C Ratio Conventional AgeBPCalibratedDateProvenience Beta-2063322820 6 40 2 26.8 % 2790 6 401020to830BCCollectednearstokemarksbetween CounterfeitersRoomandBallroom Beta-206333680 6 40 2 27.2 % 640 6 40AD1280to1410Collectedfrombreakdowninentry toStoneMountainRoom Beta-2250023750 6 50 2 26.3 % 3730 6 502290to1980BCCollectedfromfloorofBallroomat baseofwallbetweenconcentriccircle panelandwarrioranthropomorph J.F.S IMEK ,S.A.B LANKENSHIP ,A.C RESSLER ,J.C.D OUGLAS ,A.W ALLACE ,D.W EINAND AND H.W ELBORN JournalofCaveandKarstStudies, April2012 N 21

PAGE 24

beenconcentricringsthathavefadedbeyondrecognition. Nineteenindividualglyphsaresetsofconcentriccircles (Fig.3).Thereisagreatdealofvariabilityinthenumber ofcirclesinvolvedintheconcentric-circleimages.Some imageshaveonlytworings,somehavethreenestedtogether,andonecompleximagehasfourconcentriccircles ofdiminishingsize.Concentriccircleswereproducedusing twotechniques.Someimagesaredrawninblackpigment (pictographs);othersareengravedintothelimestoneof thecavewall(petroglyphs).Inthemainglyphgallery, seventeenofthecircleglyphsaredisposedinasinglearray offourverticallinesofimages(Fig.4).Theselinesinclude fromtwotosixglyphseach,spacedabout25cmapart.An arcwasdrawnpassingoveralltheconcentriccirclesinthis panel,collectingthemintoasinglecomposition.Circles, includingconcentriccircles,areoneofthemostcommon designelementsinprehistoricartinthesoutheasternUS. Theyappearinceramicdesignsandrockartoutsidecaves, andtheyarealsofrequentincaveartassemblages.Circle motifs,however,whilecommonduringtheMississippian period,arenotchronologicallydiagnosticinandofthemselves;theycanbefoundinartfromavarietyofprehistoric periods,earlyandlate. TherearesomecirclesinDunbarCavethatdohave chronologicalcontext,andthosearethethreecross-incircleimages.One,locatedawayfromthemainroom,isa simpleimagedrawninblack.Eventhissimpleversionhas Mississippianconnotations.Theothertwo(seeFig.3)are thefirstglyphsdiscoveredandaremorecomplicatedthan simplecrossesincircles.Bothareconcentriccircleimages withexteriorrayedcirclesandcrossesofdifferentforms insidetheirinnerrings.Alsonotethe1847datescratched overtheblackglyphs.Theleftimageisadenticulatecircle withastraightcrossinthemiddle.Therightimagediffers fromitscompanioninhavingatailontheouterdenticulate circleandaleft-facingswastika(BrainandPhillips,1996, p.7)inthecenter.Together,theyformaremarkablepair ofimages.Thedenticulatecrossisfoundfrequentlyin Mississippianarchaeologicalcontexts.Inparticular,itis characteristicofaMississippian-periodgorgetstylenamed aftertheCoxMoundinnortheasternAlabamawhereit wasfirstidentified(Holmes,1883).Thisgorgettypeis mostfrequentintheCumberlandRiverareaofMiddle Tennessee,however,whereitiscombinedwithimagesof woodpeckersintoadesignreferredtoastheCoxStyleby BrainandPhillips,whosuggestedthatitwasproducedby Figure3.CircleglyphsfromDunbarCave.Thesearecompleximagesthatincluderayedcirclesandinteriorcrosses.There aresimpleconcentriccirclesaswellintheglyphassemblage.Note‘‘1847’’scratchedovertheimages.Ifthisisadateengraved overtheglyphsatthattime,theglyphsmustbeolderthanthemid-nineteenthcentury. T HEPREHISTORICCAVEARTANDARCHAEOLOGYOF D UNBAR C AVE ,M ONTGOMERY C OUNTY ,T ENNESSEE 22 N JournalofCaveandKarstStudies, April2012

PAGE 25

craftspeopleintheCumberlandregion(BrainandPhillips, 1996).DunbarCaveiswithintheCumberlanddrainage, andthepresenceoftheseimageshereisstrongevidencefor aMississippianorigin. Thelasttwoimages,bothcomplexanthropomorphs, alsoimplicateMississippianartistsintheproductionof DunbarCaveÂ’sartwork.Oneimagethatweoriginallyrecordedasapairofabstractlines(Simeketal.,2007)has proved,oncloserexamination,tobetheprofileofahuman facewithfinefeaturesandlonghair(Fig.5).Thesecond anthropomorph(Fig.6a)isareclininghuman-likeform, clearlymale,withwell-definedarmsandlegs.Threefingers appearonthelowerarm.Thelowerappendages(legs) thickentowardsthefeet,andthosefeethaveclaws, suggestingananimalelementtothecreatureÂ’smakeup (Fig.6b).Thestippledheadofthisimageisquiteunusual. Thereisanaxeorcalumetabovetheheadandacurving lineextendingoutfromtheupperpartofthehead.The trunkbelowthearmsiswelldefinedandfilledwith pigment.TheanthropomorphÂ’swaistandupperlegsare coveredwithanhour-glassshapedgarment,akiltperhaps, whichhasseverallinessuggestingfoldsordecoration,but doesnotconcealtheindividualÂ’sphallus.Allofthese elementsmakeperfectsensegivenwhatweknowabout prehistoricMississippianiconography.Thaticonography focusesonthecentraltenetsofMississippianreligion, namelywarfare,death,andtheancestors(Hall,1997; Muller,1989).Mythicalwarriorswerecommonandimportantcharactersinthereligiousnarrativesandthusin Mississippianart.Theyfrequentlycombinehumanand animalcharacteristics,especiallyavianones(Brown,2007). Thesewarriorsareoftenshownwithelaboratehead decoration,includingweaponsinthehairandcurlinghair locks.KiltsarealsocommonaspectsofthewarriorÂ’s regalia.TheDunbarCaveanthropomorphisreadilyinterpretedasdepictingaMississippiancosmicwarrior, whichfitsquitewellchronologicallywiththedenticulate circlesdiscussedearlier.Interestingly,theheadofthis figure,whichistheimageintheBallroomfarthestfromthe cavemouth,pointsdown-slopetowardtheriverthatflows throughthecaveÂ’sinterior. AllofthepictographsinDunbarCaveareblack, suggestingthatcharcoalwasusedtoproducethem.However,wehaveobservedinothersitesthatpigmentswere frequentlyappliedaspaintsmadebymixingachromophore(coloringagent)withbindersinaliquidformto enhanceadhesionandpermanence(Simeketal.,2010, p.82).ToinvestigatethepigmenttechnologyatDunbar Cave,wesampledafewgrainsofpigmentfromoneofthe glyphs,alongwithafewmorefromanearbycane-torch stokemark.Acontrolsamplewasalsoobtainedfrombare limestonenearthesampledpictograph.Thisstrategywas designedtodeterminewhetherthepigmentsusedinDunbarCavehadbeenappliedwithasimplecharcoalpencil, suchastheendofacanetorch,ormorecomplexliquid paintpreparationswereapplied.UsingEDS/SEManalysis ofcomposition,wefoundthatthepictographpigment (Fig.7a)differsfromthebarelimestonecontrolsample (Fig.7b)intheelevatedpresenceofcarbon,indicatingthat thepictographswereproducedusingcharcoal.Noother Figure4.Sketchofconcentric-circlearrayenclosedbyapaintedarc.Bothpictographsandpetroglyphsareincludedinthe compositionasindicated.Approximatesize:2mwide 3 1.5mhigh. J.F.S IMEK ,S.A.B LANKENSHIP ,A.C RESSLER ,J.C.D OUGLAS ,A.W ALLACE ,D.W EINAND AND H.W ELBORN JournalofCaveandKarstStudies, April2012 N 23

PAGE 26

differenceswereobservedthatmightreflectinclusionofa paintbinder.Thissuggeststhatthepictographswere producedusingrawcharcoalpencils.Infact,thepictograph samplewasidenticaltothestokemarkincomposition (Fig.7c),signifyingthatburntcaneswereprobablyusedto producethepictographs. Oneofthemostfundamentalproblemsforstudentsof parietalrockart(pictographsandpetroglyphslocatedon walls)ischronology(Leroi-Gourhan,1971).Thepictures usuallylackassociationwithsedimentdepositscontainingtemporallydiagnosticmaterials,whetherartifactsof characteristicstyleoritemsthatcanbedatedusingradiometrictechniques.Unfortunately,thisistruefortheDunbar Cavepictographs.Aswehaveseen,thepictographsare madeofcharcoal,whichissubjecttoagingbyradiocarbon assay,butasamplefromoneofthepictures(takenwith greatcarenottodamagethevisualimage)didnotcontainsufficientcarbontoenableareading,evenbyAcceleratorMassSpectrometry.WehaveobtainedthreeAMS radiocarbonagedeterminationsfrominsideDunbarCave, allfromtorchfragmentsrecoveredfromthecavefloor (Table1).Thesesamplesweretakenfromvariousplaces insidethecave,andtheyelucidatethecomplexnatureof prehistoricactivitywithintheDunbarCavedarkzone.One calibratedagedetermination(Beta206333)coincidesperfectlywiththetime-periodwewouldexpectforthepictographiconography:betweenAD1200and1400.This samplewascollecteddeepinthecave,however,thesample mostdistantfromtheglyphs.Thesampleclosesttothe glyphs,recovereddirectlybelowthecirclepanel(Beta 225002),hasthemostancientcalibratedage,about4000 yearsbeforepresent,wellwithintheLateArchaicperiod. Thethirddetermination(Beta206332),onasampletaken attheeastedgeoftheBallroom,isWoodlandperiod, calibratedtoaroundAD900.Thus,whilepeoplewere certainlyinsidethecaveatthetimesuggestedbythepictographiconography,theywerealsotheremuchearlier.In fact,thisissomethingwefindtypicalofcavesusedprehistoricallyintheSoutheast(Simek,2010;Simekand Cressler,2001,2005,2009). 1977–78A RCHAEOLOGICAL W ORK WhiletheprehistoriccaveartinDunbarCaveisobviouslyofgreatanthropologicalinterest,thearchaeology ofthesiteisalsoquiteimportant.DunbarCaveispresently ownedandmanagedbythestateofTennessee,butthathas notalwaysbeenthecase.Someaccountsholdthatthecave wasminedforsaltpeterinthemid-nineteenthcentury,and thenitrogen-richsedimentinthecavewasalsominedfor fertilizerinthelatenineteenthcentury.Thecavethen beganalonghistoryasatouristattraction.Overtime,the sitewasdevelopedasaresortbyseveralinterests,including RoyAcuffafterWWII,anditsawtheinstallationofgolf andswimmingfacilities,musicalattractionsbefittingasite soclosetoNashville,andarecreationallake(VanWest, 1998;Whidby,1999;Matthews,2005).Oneeffectofthese ratherspecializeduseswastheprotectionandeventual sealingofvestibuledepositsatthemouthofthecave.First woodenandlaterconcreteplatformswereconstructedon topofthesedimentsthatfilledthelargerockshelter protectingthecaveentrance(Butler,1977).Thus,where somanyofthegreatsheltersitesintheeasternwoodlands werelootedfortheirartifacts,DunbarCave’sarchaeology remainedintactforthemostpart.In1973,thestate acquiredthesite,anditsprotectionwasguaranteed. WhenthestateofTennesseetookpossessionofDunbar Cave,testexcavationswereinitiatedbothinsideand outsidethecavebytheStateDivisionofArchaeology (DOA)todeterminethenatureandextentofthesite’s archaeologicalrecord.DOAexcavationswereconducted ontwoseparateoccasions.BrianButlerdirectedavery smallarchaeologicalprojectin1977todeterminewhether intactarchaeologicaldepositswerepresentandhowdeeply theywereburiedbeneaththefill(Butler,1977).Threetest pitswereexcavatedattheentranceofthecave,after breachingtheconcretedance-floorslab,andthreeadditionalpitswereplacedalongthepathjustinsidethecave. Thethreetestpitsoutsidethecaveyieldedcharcoal,bone, shell,lithics,andpotteryindicatinghumanoccupation. Figure5.Pictographofananthropomorphheadandface inprofile. T HEPREHISTORICCAVEARTANDARCHAEOLOGYOF D UNBAR C AVE ,M ONTGOMERY C OUNTY ,T ENNESSEE 24 N JournalofCaveandKarstStudies, April2012

PAGE 27

Thosetestpitswithinthecaveinteriorexhibitedonlya minimalamountofculturalmaterial. In1978,RobertPaceandVictorHoodledalarger, moreinvolvedarchaeologicalexcavationundertheauspicesoftheDOA(PaceandHood,1978).Forthisinvestigation,theteamdugatrenchmadeupofeight2 3 2meter unitsrunningnorth-southatthemouthofthecave (Fig.8).Theunitswereexcavatedinarbitrary10cmlevels startingfromanestablisheddatum,withverticalcontrol beingmaintainedbyatransit.Thedepositswerefoundto bemorethan5.5m(nearly20ft)deepinsomeareas, makingthisanextraordinaryarchaeologicaldepositwith technicalchallengesforitsexcavators.Morethansixty storageboxesofmaterialwerefilled,acollectionthat presentlyresidesattheUniversityofTennessee.Diagnostic artifactsspantheperiodfromLatePaleo-Indian(about 10,000yearsago)totheMississippian(toAD1500),and theyincludeArchaic(10,000to3000yearsago)and Woodland(3000to1000yearsago)culturalmaterials.Lithic artifacttypesinroughsequenceincludeaBeaverLakePoint, Kirkclusterpoints,LittleRivertypes,Ledbettertypes, numerousstraight-stemmedpointtypes,andHamiltonand Madisonprojectilepoints.Woodlandceramicscomprise variouslimestone-temperedforms,allinlowquantities,and cord-markedlimestone-temperedwaresintheuppermost Woodlandlayers.Shell-temperedceramicsbearwitnesstoa richMississippianpresenceatthetopofthedeposit.Given thischronologicalspan,theDunbarCavesequenceisas Figure6.Pictographofareclininganthropomorphwithanimalfeatures.A.Photographofthepictograph.B.Sketchofthe pictographbasedonfieldobservationusingtheredportionofthevisible-lightspectrum. J.F.S IMEK ,S.A.B LANKENSHIP ,A.C RESSLER ,J.C.D OUGLAS ,A.W ALLACE ,D.W EINAND AND H.W ELBORN JournalofCaveandKarstStudies, April2012 N 25

PAGE 28

completeasanyineasternNorthAmerica(PaceandHood, 1978).However,wedonotdiscusstheartifactsinmoredetail herebecause,asnotedbelow,therearegeneralproblems withproveniencestemmingfromthecomplexstratigraphyof thesiteandthewayitwasexcavated. Units96N0through106N0(seeFig.8)appeartohave beenexcavatedsimultaneously,whilethetoplevelsof 108N0and110N0werestrippedwithashovelafterbeing identifiedascontainingfill.Unitswereexcavatedtodifferentdepths.Forexample,Unit96N0wasexcavatedto 2.65mdeep,wherelargeboulderspreventedanymore excavation.Limestonebreakdownpreventedexcavation deeperthan2.75minUnit98N0.Unit100N0wasexcavated toadepthof2.65mandcontainednodiagnosticartifacts. Thequantityofarchaeologicalmaterialpresentinunits 102N0through106N0necessitateddeeperexcavationsto documenttheprehistoricuseofthecaveinitsentirety. Unit102N0wasexcavatedtoadepthof4.45m.Ceramics, lithics,andbonewererecoveredthroughtheentiredeposit. Theartifactsequenceinthesquare,however,isconfusing. Withregardtothepottery,manylevelscontainbothlimestone-andshell-temperedpottery.Thisisunusualbecause limestonepotterytendstobeearlier,associatedwiththe Woodlandperiod,andshelltemperisalaterMississippian characteristic(LewisandKneberg,1995).Evenmore problematicarelevelscontainingshell-temperedpottery thatliebeneathlevelsidentifiedasArchaicfromdiagnosticlithics.Therewouldappeartobeproblemswiththe stratigraphyinthisunit. FromUnit104N0,onlytendiagnosticartifactswere recovered.Theunitwasexcavatedtoadepthof3.35m. Shell-temperedshardswerefoundinlevels2and4, convenientlyabovelimestone-temperedshardsthatwere recoveredfromlevels4,5,8,and17.Theonlydiagnostic lithicsweretwoLateArchaicstemmedpointsfromlevel8, Figure8.Planmapof1978vestibuleexcavationsshowing2 3 2munitlocationsanddesignations,after1978fieldnotes byPaceandothers. Figure7.ElementspectrographsshowingresultsofEDS/ SEManalysesofmaterialcomposition.A.Analysisof pictographpigment.B.Analysisofbarelimestonewallasa control.C.Analysisofcane-torchstokemark. T HEPREHISTORICCAVEARTANDARCHAEOLOGYOF D UNBAR C AVE ,M ONTGOMERY C OUNTY ,T ENNESSEE 26 N JournalofCaveandKarstStudies, April2012

PAGE 29

sothisunithasprobableWoodlandpotterystratigraphicallybelowArchaicprojectilepoints.Again,stratigraphic problemsareapparent. NumerousartifactswererecoveredfromUnit106N0. Intheupperlevels,sketchesweremadeofthevarious deposits.Thiswastoaccountforthelackofthesediment beingscreenedbecauseitwasconsideredfillanddiscarded. Levels2through5wereremovedbyshovel,astheywere entirelycomposedoffill.Hand-excavationbeganonlywith level15atadepthof2.95m,wheretherewasnointrusion fromthered-clayfill.Excavationcontinuedtolevel31, seekingthebottomoftheculturaldeposits.Ataround 5.2m,thedecisionwasmadetouseabackhoetofacilitate digging.At5.55m(level41),waterbegantoimpedethe excavation.Verticalcontrolwasmaintainedforthreemore levelsbeforeitbecameimpossibletounderstandthe contextoftheartifacts.Theproblemswithstratigraphic controlobservedforUnit102NOareevenmoreapparent whentherichassemblagesfrom106NOareexamined. Withregardtopottery,MississippianandWoodland shardsareroutinelyfoundtogetheruntillevel29(witha maximumdepthof4.45m).Twoshell-temperedshardsare foundbelowthisinlevel38(5.25–5.35m). Becausetherelationshipbetweenthearchaeological levelsdefinedduringexcavationandthetemporally diagnosticartifacttypesisconfused,wemadeaneffort toresolvetheconfusionbysecuring 14 Cagedeterminations fromseveral1978excavationunits.Table2showsthe twelvedatesobtainedandthemeasurementandproveniencedataassociatedwiththesamples.Allanalyseswere carriedoutattheUniversityofTennesseeRadiocarbon Laboratory.Thefirstfourdeterminationsweremadeon samplesfromUnit102NO.Whilepartiallyconformingto whatmightbeexpectedgivendepthsbelowsurface,the twoupperdatesareinverted,indicatingstratigraphic mixingevenatthisfinescaleofanalysis.Thesecondset offourdeterminationsisfromUnit106NO.Thefirstpair isinvertedfromstratigraphicexpectations,whilethe secondtwo,supposedlyassociatedwithbisonremains, areolderthantheirassociationwouldindicate.Thelast fouragesarefromUnit106N2W.Thefirstpair,in sequencewithinasinglehearthfeature(F11),isstratigraphicallyinverted,whilethelasttwodeterminationsare statisticallyidenticaleventhoughtheycomefromsuperimposedlevels.Thus,new 14 Cagedeterminationsdonot clarifythechronology.Instead,ageinversion,inaccurate association,andlackofagreementbetweenstratigraphic positionandradiometricagecharacterizetheDunbarCave vestibuleexcavationsequence. Clearly,thereareproblemswiththestratigraphyfrom theentire1978excavation.Thismightbeduetooneor bothoftwofactors.Itmaybethatthesiteisgreatlydisturbedandhasloststratigraphicintegritybecauseofpostdepositionalprocessesaffectingthesediments,something thatisalwayspossibleinthecomplex,dynamicdeposits thatfillcavesandrockshelters.Or,itmaybethat excavationstrategiesfailedtoestablishaccurateproveniencefortheartifactsrecoveredfromthesite.Webelieve thatthelatterproblemexplainsobservedincongruitiesin theDunbarCaveartifacts. AfterexcavationatDunbarwascompletedandthe deeptrenchat106Nexpandedtotwounitswidetoprovide aprofile,thedrawingshowninFigure9wasmadeofthe deepsequence.Therearecomplexbutclear,perceptible stratigraphiclayersandlimitsrecordedbytheartist.What isalsoapparentisthatthearchaeologicallayersarenotflat andlevel;indeedtheyaremoundedwithslopingsides,as mightbeexpectedforatalus-slopedeposit,andthis characterisamplifiedthedeeperintothedepositonegoes. Excavationstrategy,however,wastobringalleightunits toasinglehorizontalplanepriortoexcavationwith absolutedepthbelowthesitedatumdeterminingthe stratigraphiccollectionunit.Accordingtothenotebook, ‘‘Level1wasdesignedtobringtheentiretrenchtoa uniformhorizontalplaneatthelowestpointinthe trench.…allotherlevelswereuniformly10cminthickness.’’Thus,level2wasdefinedasaten-centimeterspitin eachsquarebetween1.65and1.75centimetersbelowthe surface.Alllevelshadthesameverticaldepthregardlessof unitorsedimentarycontext.Inessence,thecollectionunits wereconstantvolumetricallybutstratigraphicallyarbitrary.Theeffectofthiswasdevastatingforartifact provenience.Take,forexample,asinglecollectionspitin Unit106NOwhere,accordingtothedrawnprofile, seventeenidentifiablestrataaremixedtogetherintoone horizontal10cmunitlevel(seeFig.9).Thisexampleis replicatedtimeandagainacrossthetrench.Thus,itisnot surprisingthatartifactsfromdifferenttimeperiodsoccur innearlyeverylevelfromeverysquare.Thebadnewsis thatwecannotrecoverstratigraphicprovenienceforthese materials.Thegoodnewsisthatthesiteitselfprobablyhas intactstratathatcanultimatelyyieldcontrolleddatato newexcavation. Therearesomeaspectsofthe1978collectionthatcan stillprovideuswithimportantinformation.Weinand,for example,hasundertakenastudyofthearchaeofaunas fromthe1978DunbarCaveexcavations.Theassemblage yieldsveryfewsurprises.Speciespresentincludedeer, smallmammals,fish,turtles,andturkeys,alltypicalfoodpreyspeciesfoundinsoutheasternarchaeologicalsites. Twoflyingsquirrelsandaswan,morerareinarchaeaofaunas,havealsobeenidentified.Themajorityofthe DunbarCavebonesfromalllevels,includingthedeepest, areburned,indicatingahumanroleintheiraccumulation. Somearefashionedintotools(awlsorpunches),and butcherycut-marksareevidentonsomeremains. Thebonesofthreeparticularlyinterestingspecieswere foundinunits108Nand106N.Oneofthemwasidentified in1978asoneoftheveryfew Bisonbison bonesever identifiedinaTennesseearchaeologicalsite.Infact,atotal ofthreebuffaloboneshavenowbeenidentified,allfrom thesameprovenience.Thebonesincludeametacarpal,first J.F.S IMEK ,S.A.B LANKENSHIP ,A.C RESSLER ,J.C.D OUGLAS ,A.W ALLACE ,D.W EINAND AND H.W ELBORN JournalofCaveandKarstStudies, April2012 N 27

PAGE 30

identifiedin1978,whichshowsclearaffinitytobison ratherthancattle(Fig.10).Inaddition,bothafirstanda secondphalanxarticulatewiththemetacarpal:MNIis therefore1.NocutmarksarepresenttoshowthatDunbar Caveinhabitantswerehuntingorbutcheringbison.The chronologicalpositionofthisanimalisobviouslyofgreat interest,butunfortunately,wecannotrelyonstratigraphic associationtoprovideanindicationofage.Thereforea sectionofbone,approximately1cm 3 3cm,wasremoved fromthedistal,posteriormedialsurface.Thelocationfor boneremovalwaschosentominimizedamagetothebone andtoavoidareascontainingmeasurementlandmarks,in casetheboneissubjectedtofuturestudy.Thesectionwas submittedtoBetaAnalytic,Inc.,forAMSdatingofthe collagenfraction.Anagedeterminationwasreturnedof 1455 6 45BP.Thecalibratedageofthesampleindicates thatthebisondiedbetweenAD1420and1490(calibrated at2 s withINTCAL04[Reimeretal.,2004]).Thisconfirms aMississippianassociationfortheanimalandrepresents anearlydateforthisspeciesintheMiddleTennessee region.Thisdeterminationdoesnotoverlapwiththe earlieragesforsupposedlyassociatedcharcoaldiscussed above(Table2). Teethandseveralmandiblefragmentshavebeen identifiedaselk( Cervuselaphus ).Thesealsocomefrom Units106Nand108N,althoughMNIstillis1.Incontrast tothebison,theelkmandiblefragmentsdoshowcut marks,suggestinghumanutilizationofthislargeartiodactyl.Therearealsoremainsofthelefthumerusfromalarge blackbear( Ursusamericanus )fromUnit108Nthatshow evidenceofrodentgnawing,indicatingthatitwasexposed forsometimepriortoburial.Chronologicalcontrolfor theelkandbearislacking,andtheiragesremaintobe determined. Table2.RadiocarbonagedeterminationsfromvestibuleexcavationsatDunbarCave(40MT43).Samplesineachunitlistedin deepeningstratigraphicorder.AllassaysperformedbyUniversityofTennesseeRadiocarbonLaboratoryusingLiquid ScintillationCountingandcalibratedat2 s withINTCAL04(Reimeretal.,2004).Locationdescriptionsinthetablefollow (PaceandHood,1978). Sample Reference Measured AgeBPCalibratedDate Archaeological Level DepthBelow Surface,cmProvenience Unit102NO UT08-0513870 6 782569to2064BC5102Singlepieceofcharredwood fromgeneralmidden UT08-0503770 6 702459to1984BC10102.52Possiblyresidualfeature-no depth-pieceofantlerin association UT-08-0485570 6 1074690to4174BC21103.61Concentratedscatterof woodcharcoal UT-08-0436125 6 1415371to4718BC21103.7Charcoalscatter 10cmin diameterNhalfofcentral portion Unit106NO UT06-0344300 6 703309to2668BCF.5S1/2103.45Charcoalinashylayer directlyaboveFeature5westprofile UT06-0313900 6 702572to2151BCF.5S1/2103.72Woodcharcoaldirectlybelow firedclaycapofFeature5 UT08-052960 6 81AD897to1251 ?????? Charcoalina(hearth?)associatedwiththe Bison bone UT08-052(2)1000 6 93AD782to1224 ?????? Charcoalina(hearth?)associatedwiththe Bison bone Unit106N2W UT06-0286740 6 705752to5521BCF.11 ??? CharredwoodLevelIV, Feature11 UT08-0536700 6 735724to5490BCF.11 ??? CharcoalW1/2F.11,Lev.7 UT08-0266740 6 1435980to5386BC37 ??? Woodcharcoalgatheredfrom scatterthroughoutlevel UT08-0256790 6 1095966to5511BC42 ??? HandsortedcharcoalfromconcentratedscatterinNE1/4 T HEPREHISTORICCAVEARTANDARCHAEOLOGYOF D UNBAR C AVE ,M ONTGOMERY C OUNTY ,T ENNESSEE 28 N JournalofCaveandKarstStudies, April2012

PAGE 31

R ECENT A RCHAEOLOGICAL W ORK AtthetimetheprehistoricartinDunbarCavewas discovered,thecavewasprotectedwithanoldgatethatcould easilybebreeched.InJulyof2006,ateamofTDEC personnelandvolunteersledbyindependentconservationist andgate-builderKristenBoboinstalledanew,more substantial,andbat-friendlygate.Onceitwaswell-defended, theprehistoriccaveartinDunbarCavewasannouncedwith apublicceremonyatthecave.DunbarCaveisthefirst Tennessee,indeedfirstNorthAmerican,dark-zonecave-art siteopentotheinterestedpublic,andthecaveartispartof theinterpretiveprogramofferedtovisitors.Thebat-friendly gatehashadtheintendedconsequenceofencouragingnew bathabitation.Unfortunately,whitenosesyndromewas recognizedamongtheDunbarCavebatpopulationinMarch 2010,andthecavewasclosedtoalltraffic. InMayof2007,asDunbarCaveStateNaturalArea preparedinterpretiveexhibits,weundertooksmall-scale testexcavationsintheglyphchamberofthecave.Two1 3 0.5munitswereexcavatedintheBallroom,onenearthe warriorimageandatthespotwhereaLateArchaic(i.e., Ledbetter)projectilepointhadbeenfoundjustbelowthe surfacenearthecavewall(Fig.11a),andtheotherin seeminglyintactdepositsattheoppositewallofthe chamber.Theseunitsattheedgesoftheroomwerechosen becausethedepositsinthecenterhadbeensubstantially modifiedbyconstructionofavisitorpathway.Stratigraphyinthetworoom-edgeunitswasessentiallythesame: Level1(8to10cmthick):Bandeddarkandlightfine water-laidsilts,containshistoricartifacts(wood,glass, nails),heavilytrampled. Level2(1to2cm):Solidtravertineflooroverentireunit composedofbandedlayersofcalciumcarbonate, archaeologicallysterile. Level3(10to50 + cm):Homogeneousdarkreddishbrown sandysiltwithoccasionalpla quettesandblocks,archaeologicallysterile. Interestingly,veryfewprehistoricartifactswereuncoveredduringthesetestexcavations,despitetheuseof quarter-inchmeshtoscreenallsediments,butthosethat werefoundarerelativelylargeandaremadeofhighqualitychert.Theyincludethestemmedpointshownin Figure11aandthebifacefragmentshowninFigure11b. Ineverycase,theseartifactswerelyingdirectlyincontact withthetopofLevel2,thetravertinesurface. InMay2008,thestateinstalledthreeinterpretivepanels infrontoftheprincipalglyphwallsintheBallroom,each Figure9.Profileof1978excavationtrenchatUnit106N,after1978fieldnotesbyPaceandothers.Shadedareashowssingle 2m 3 10cmhorizontalexcavationspitcontainingseventeenstrata. J.F.S IMEK ,S.A.B LANKENSHIP ,A.C RESSLER ,J.C.D OUGLAS ,A.W ALLACE ,D.W EINAND AND H.W ELBORN JournalofCaveandKarstStudies, April2012 N 29

PAGE 32

supportedbytwopostsburiedinthecavesediments.We wereaskedtoexcavatetheholesforthesupportlegs,in effect,providingsixmoreshoveltestsintheareaofthe mainglyphconcentrations.Eachoftheseunitswas20cm squareandexcavatedtoadepthofabout50cmwhere possible.Intwocases,rockswereencounteredatmore than30cmbelowthesurface.Withoneexception,allsix shoveltestsdisplayedthesamestratigraphyasthetestunits describedearlier,althoughinthenorthernmosttestunit, closesttothewarriorpanel,asomewhatmorecomplex sequencewasencounteredinandbelowLevel3.Webriefly describethatprofileasrepresentativeofallsix2008units. Atthetop,10cmofbandeddarkandlightfinewater-laid sedimentswasencountered,Level1asdefinedin2007.In allsixunits,alayercorrespondingtoLevel2,thetravertine surface,wasidentifiednext,althoughinfourunitsLevel2 wasdecomposed,reflectedbytravertineclastsandcalcite sand.Level2variedfrom1to3cminthickness.Level3, 20cmthick,consistedofdark,reddish-brownsandyloam withoccasionallensesofsiltandclay.Level4,an8cm layerofplaquettesrichincalcite,andLevel5,lightcolored,cross-beddedsiltsandsands,decends50cmfrom thebaseofLevel4.Onlytwoartifactswereencountered duringexcavationofthesixpostholes:awhitehistoric ceramicshardatthetopofLevel1,andalargechertcore (Fig.11c)lyingonthetravertinesurfaceatthecontact betweenLevels1and2. Eventhelimitedinformationtheseexcavationsprovide concerningprehistoricuseoftheinteriorofDunbarCave isintriguing.Itseemsthattherewas,atsometimeinthe past,atravertinesurfacethatextendedacrosstheBallroom notunlikethosepresenttodayinotherpartsofthecave, includingthenearbyCounterfeiter’sRoom.Thatsurface, aselsewhere,probablyheldwaterandmayhavebeen composedofgoursorotherwater-relatedspeleothems.All oftheexcavatedprehistoricartifactsareLateArchaicin aspectandlaydirectlyonthesurfaceofthatsubmerged floor.Recallalsothatthesingle 14 Cagedetermination fromthisroomisLateArchaic.Wespeculatethatthese artifactswereintentionallydeposited,‘‘offered’’intothe watersoftheBallroomsome4000yearsagobyArchaic periodvisitors,longbeforetheartworkwasaddedtothe walls.SuchpracticesarewellknowninMesoamerica, whereritualsandofferingswerecommonlymadeincaves, ashasbeendemonstratedbothethnographicallyand archaeologically(seepapersinBradyandPrufer,2005). Suchapracticewouldexplainatleastsomeevidencefor ArchaiccaveuseintheSoutheastaswell(Simekand Cressler,2009;Simeketal.,1998;Crothersetal.,2002). C ONCLUSIONS DunbarCaveisaremarkableprehistoricsite,notjust forTennesseebutfortheentireEasternWoodlands. Humanuseofthecaveinteriorspansatleast4000years. Peoplewereinterredinthecave,cavemineralsmayhave beenprocured,andthecaveinteriorwasextensively explored.Aprehistoricsanctuarywascreatedduringthe Mississippianperiod(aroundAD1300)withdarkzone, parietalcaveartdepictingvariousiconographicimages, includingheroiccharactersfromcosmologicalnarratives. ItisevenpossiblethatArchaicvisitors(about2000BC) expressedtheirreverencefortheunderworldbyleaving large,finely-workedstonetoolsinpoolsinthecave interior.Outsidethecave,thevestibulerockshelterwas periodicallyoccupiedbeginninglateinthePleistoceneand continuingthroughallHoloceneprehistoricphasesinthe Southeast,withbothhuntingandagriculturalpeoples repeatedlyvisitingthesitetohuntitssurrounds,toobtain stonefortheirtools,toburytheirdead,andtocarryout religiousrituals. ArchaeologicalworkatDunbarCavewillcontinue,and thereareanumberofaspectsyettobeexamined.For example,avisittothehilltopcrowningDunbarCavein 2007revealedthatinmanyplacestherockyoutcropshave beenexcavated,andinthoseareaslargeandfine-grained chertnoduleswereapparent,sometestedwithhammerstoneblows.ThestaffofDunbarCavewasawareofsome ofthese,butinashortperiodoftimeweidentifiedatleast ahalf-dozenmajorchertquarrysitesoutsidethecave. Figure10.ComparisonofDunbarCavebisonmetacarpalto modernbisonandcow. T HEPREHISTORICCAVEARTANDARCHAEOLOGYOF D UNBAR C AVE ,M ONTGOMERY C OUNTY ,T ENNESSEE 30 N JournalofCaveandKarstStudies, April2012

PAGE 33

Someofthesearequiteextensive,andinsomeplaces numerouschertnodulesderivedfromtheoutcropsare piledinassociationwithknappingdebris.Thevicinityof DunbarCavewasprobablyasignificantprehistoricchert minelocatedneartherenownedDover,Tennessee,chert quarries.In2006,wevisitedapartofthecaveinteriorwe hadneverseenandobservedseverallocalitieswhere gypsumcrustshadbeenbatteredfromthecavewallsand removed.Ahammerstonelayontheflooratthebaseof oneofthesewalls.Itmaybethatprehistoricminerswere activebothinsideandoutsideDunbarCave,astheywere inseveralTennesseesites.Wehaveyettoundertake intensiveworkonthesediscoveries. Thereis,thus,agreatdealofresearchyettodoat DunbarCave.Thecavehasmorethaneightmilesof mappedpassages,butwehavevisitedlessthanaquarterof thatlength.Thereisaninternalarchaeologicalsurveyto carryout,seekinganddocumentingartandotheractivities likeminingandburialthatmighthavetakenplaceinthe caveÂ’srecesses.Chertminingoutsidethecaveshouldalso beexaminedindetail,andtheconsiderableremaining vestibulesedimentswarrantfurtherinvestigation.Dunbar Caveisaremarkablearchaeologicalresource.Itdeserves andwillreceivetheattentionrequiredbyitscomplex, multifacetedandlong-termprehistoricuse. A CKNOWLEDGEMENTS TheauthorsthankthestaffofDunbarCaveState NaturalArea,especiallySuperintendentBobWells.The Figure11.ArtifactsrecoveredonthesurfaceofburiedspeleothemfloorintheBallroomofDunbarCave.A.LateArchaic (Ledbettertype)projectilepoint.B.Bifacefragmentfoundintestexcavation.C.Corefragmentfoundinpostholeexcavation. J.F.S IMEK ,S.A.B LANKENSHIP ,A.C RESSLER ,J.C.D OUGLAS ,A.W ALLACE ,D.W EINAND AND H.W ELBORN JournalofCaveandKarstStudies, April2012 N 31

PAGE 34

FriendsofDunbarCaveandtheentirecommunityof Clarksvillehavebeenverysupportiveofarchaeological researchatthecave.JerryClark,whodiscoveredseveralof theglyphs,andRonnieHunterdeservespecialthanks. SarahC.Sherwood(Sewanee,theUniversityoftheSouth) helpedusduringthe2007testexcavationsandadvised HeatherWelbornonherstratigraphicresearchinthesite. TennesseeStateArchaeologistsNickFielderandMike Mooreprovideduspermitstodothework,andmadethe 1977–78collectionsavailabletous.BillLawrenceandJohn Froeschauer,archaeologistswiththeTennesseeStateParks, continuetheirgreatsupport.StuartCarrollhelpedusinthe gatingprojectandwas,asalways,instrumentalinsustaining generalmorale.PattyJoWatsonmadesubstantiveimprovementstothepresentationandgreatlyimprovedthe clarityinherreviewofthispaper;asecondanonymous reviewerwasalsohelpful.Fundingfortheresearch describedherecamefromtheNationalScienceFoundation, theUniversityofTennessee,andanonymousdonorstothe UTCaveArchaeologyResearchTeam. R EFERENCES Boszhardt,R.F.,2003,DeepCaveRockArtintheUpperMississippi Valley:St.Paul,MI,PrairieSmokePress,94p. Brady,J.E.,andPrufer,K.M.,2005,IntheMawoftheEarthMonster: StudiesofMesoamericanRitualCaveUse:Austin,Universityof TexasPress,448p. Brain,J.P.,andPhillips,P.,1996,ShellGorgets:StylesoftheLate PrehistoricandProtohistoricSoutheast:Cambridge,Massachusetts, PeabodyMuseumPress,542p. Brown,J.A.,2007,OntheidentityofthebirdmanwithinMississippian periodartandiconography, in ReillyIII,F.K.,andGarber,J.F., eds.,AncientObjectsandSacredRealms:InterpretationofMississippianIconography:Austin,UniversityofTexasPress,p.56–106. Butler,B.M.,1977,ApreliminaryarchaeologicalassessmentofDunbar Cave(40MT43):Nashville,DivisionofArchaeology,Tennessee DepartmentofConservation. Creswell,B.A.,2007,PhaseIArchaeologicalSurveyofCavesand RockshelterswithintheProposedCorridoroftheKnoxvilleParkway (SR475)inAnderson,Knox,andLoudonCounties,Tennessee: Knoxville,UniversityofTennesseeArchaeologicalResearchLaboratories. Crothers,G.,Faulkner,C.H.,Simek,J.F.,Watson,P.J.,andWilley,P., 2002,Woodlandperiodcaveuseintheeasternwoodlands, in Anderson,D.G.,andMainfort,R.C.,Jr.,eds.,TheWoodland Southeast:Tuscaloosa,UniversityofAlabamaPress,p.502–524. Diaz-Granados,C.,andDuncan,J.R.,2000,ThePetroglyphsand PictographsofMissouri:Tuscaloosa,UniversityofAlabamaPress, 280p. Diaz-Grana dos,C.,Rowe,M.W.,Hyman,M.,Duncan,J.R.,and Southon,J.R.,2001,AMSradiocarbondatesforcharcoalfromthree Missouripictographsandtheirassociatediconography:American Antiquity,v.66,p.481–492. Faulkner,C.H.,ed.,1986,ThePrehistoricNativeAmericanArtofMud GlyphCave:Knoxville,UniversityofTennesseePress,124p. Faulkner,C.H.,1988,Astudyofsevensoutheasternglyphcaves:North AmericanArchaeologist,v.9,no.3,p.223–246. Faulkner,C.H.,Deane,B.,andEarnest,H.H.,Jr.,1984,AMississippian periodritualcaveinTennessee:AmericanAntiquity,v.49,no.2, p.350–361. Faulkner,C.H.,andSimek,J.F.,1996,Mudglyphs:Recentlydiscovered caveartineasternNorthAmerica:InternationalNewsletteronRock Art,v.15,p.8–13. Hall,R.L.,1997,AnArchaeologyoftheSoul:NorthAmericanIndian BeliefandRitual:Champaign,UniversityofIllinoisPress,222p. Holmes,W.H.,1883,ArtinshelloftheancientAmericans, in 2ndAnnual ReportoftheBureauofAmericanEthnology,1879–80:Washington, D.C.,GovernmentPrintingOffice,p.179–305. Leroi-Gourhan,A.,1971,Pre histoiredel’ArtOccidental,secondedition: Paris,Mazenod,499p. Lewis,T.M.N.,andLewis,M.D.K.,1995,ThePrehistoryofthe ChickamaugaBasininTennessee:Knoxville,UniversityofTennessee Press,twovols.,683p. Matthews,L.E.,2005,DunbarCave:TheShowplaceoftheSouth: Huntsville,Alabama,NationalSpeleologicalSociety,145p. Muller,J.,1989,Thesoutherncult, in Galloway,P.K.,ed.,The SoutheasternCeremonialComplex:ArtifactsandAnalysis:Lincoln, UniversityofNebraskaPress,p.11–26. Pace,R.A.,andHood,V.P.,1978,ArchaeologicalInvestigationsat DunbarCave:AStratifiedRockshelter:Nashville,Divisionof Archaeology,TennesseeDepartmentofConservation. Reimer,P.J.,etal.,2004,IntCal04terrestrialradiocarbonagecalibration, 0–26calkyrBP:Radiocarbon,v.46,no.3,p.1029–1058. SaboIII,G.,2008,Rockartandthestudyofancientreligionsin southeasternNorthAmerica, in Fogelin,L.,ed.,Religion,Archaeology,andtheMaterialWorld:Carbondale,CenterforArchaeological Investigations,SouthernIllinoisUniversity,occasionalpaper36, p.279–296. Simek,J.F.,2010,Afterword:Onwardintothedarkness(stillfollowing thelightofPatWatson’slamp,ofcourse), in Dye,D.,ed.,Cave ArchaeologyintheEasternWoodlands:EssaysinHonorofPattyJo Watson:Knoxville,UniversityofTennesseePress,p.261–270. Simek,J.F.,Blankenship,S.,Herrmann,N.,Sherwood,S.C.,and Cressler,A.,2010,NewcaveandrockartsitesinTennessee:2007, in Baumann,T.,andGroover,M.,eds.,Pottery,Passages,Postholes, andPorcelain:EssaysinHonorofCharlesH.Faulkner:Knoxville, McClungMuseum,ReportofInvestigationsseries,Universityof Tennessee,p.71–88. Simek,J.F.,andCressler,A.,2001,Issuesinthestudyofprehistoric southeasterncaveart:MidcontinentalJournalofArchaeology,v.26, no.2,p.233–250. Simek,J.F.,andCressler,A.,2005,Imagesindarkness:Prehistoriccave artinsoutheasternNorthAmerica, in Loendorf,L.,Chippendale,C., andWhitley,D.,eds.,DiscoveringNorthAmericanRockArt: Tucson,UniversityofArizonaPress,p.93–113. Simek,J.F.,andCressler,A.,2009,PrehistoriccaveartinsoutheasternNorthAmerica, in White,W.B.,ed.,Proceedingsofthe15th InternationalCongressofSpeleology:Huntsville,Alabama,National SpeleologicalSociety,p.135–139. Simek,J.F.,Douglas,J.C.,andWallace,A.,2007,Ancientcaveartat DunbarCaveStateNaturalArea:TennesseeConservationist,v.23, no.5,p.24–26. Simek,J.F.,Franklin,J.D.,andSherwood,S.C.,1998,Thecontextof earlysoutheasternprehistoriccaveart:Areportonthearchaeologyof 3rdUnnamedCave:AmericanAntiquity,v.63,no.4,p.663–675. VanWest,C.,1998,DunbarCaveStateNaturalArea, in VanWest,C., ed.,TennesseeEncyclopediaofHistoryandCulture:Nashville, Tennessee,RutledgeHillPress,266p. Wagner,M.J.,1996,Writteninstone:Anoverviewoftherockartof Illinois, in Faulkner,C.H.,ed.,RockArtoftheEasternWoodlands: SanMiguel,California,AmericanRockArtResearchAssociation occasionalpaper2,p.47–79. Waring,A.J.,Jr.,andHolder,P.,1945,Aprehistoricceremonialcomplex inthesoutheasternUnitedStates:AmericanAnthropologist,v.47, no.1,p.1–34,doi:10.1525/aa.1945.47.1.02a00020. Whidby,J.,1999,IdahoSpringsandDunbarCaveResort:Theshowplace ofthesouth:JournaloftheCumberlandSpeleanAssociation,v.6, no.1,p.7–12. T HEPREHISTORICCAVEARTANDARCHAEOLOGYOF D UNBAR C AVE ,M ONTGOMERY C OUNTY ,T ENNESSEE 32 N JournalofCaveandKarstStudies, April2012

PAGE 35

CANDIDATECAVEENTRANCESONMARS G LEN E.C USHING U.S.GeologicalSurvey,AstrogeologyScienceCenter,2255N.GeminiDr.,Flagstaff,AZ86001,USA,gcushing@usgs.gov Abstract: ThispaperpresentsnewlydiscoveredcandidatecaveentrancesintoMartian near-surfacelavatubes,volcano-tectonicfracturesystems,andpitcratersanddescribes theircharacteristicsandexplorationpossibilities.Thesecandidatesareallcollapse featuresthatoccureitherintermittentlyalonglaterallycontinuoustrench-like depressionsorinthefloorsofsheer-walledatypicalpitcraters.Asviewedfromorbit, locationsofmostcandidatesarevisiblyconsistentwithknownterrestrialfeaturessuchas tube-fedlavaflows,volcano-tectonicfractures,andpitcraters,eachofwhichformsby mechanismsthatcanproducecaves.Althoughwecannotdeterminesubsurfaceextentsof theMartianfeaturesdiscussedhere,somemaycontinueunimpededformanykilometers ifterrestrialexamplesareindeedanalogous.Thefeaturespresentedherewereidentified inimagesacquiredbytheMarsOdyssey’sThermalEmissionImagingSystemvisiblewavelengthcamera,andbytheMarsReconnaissanceOrbiter’sContextCamera.Select candidateshavesincebeentargetedbytheHigh-ResolutionImagingScienceExperiment.Martiancavesarepromisingpotentialsitesforfuturehumanhabitationand astrobiologyinvestigations;understandingtheircharacteristicsiscriticalforlong-term missionplanningandfordevelopingthenecessaryexplorationtechnologies. I NTRODUCTION SinceOberbecketal.(1969)firstproposedthatsome lunarrillesmightbecollapsedlavatubes,theexistence, characteristics,andpotentialutilityofextraterrestrialcave systemshavebeenextensivelydiscussed,particularlyfor theMoonandMars(e.g.,Horz,1985;Bostonetal.,2003). Consideringthatbasalticvolcanismisgenerallyanalogous betweentheEarthandMars(e.g.,Glazeetal.,2005), volcaniccavesarebelievedtobefairlycommononMars (CruikshankandWood,1972;Horz,1985),butuntil recently,technicallimitations,suchasthespatialresolution,arealcoverage,andviewingperspectiveoforbiting instrumentshavehinderedtheirdetection.Bleacheretal. (2007a,2007b)identifiednumeroustube-fedlava-flow systemsontheflanksofOlympusMonsandelsewherein theTharsisregionofMarsandinferredsomecollapsed lava-tubesectionstobeskylights.However,theyalso suggestedthatthehostlavatubeshadcompletelycollapsed anddidnotelaborateuponthepossibilityofintactcave systems.Thispaperpresentsnewobservationsofcandidate entrancesintonear-surfacelavatubesandvolcano-tectonic caves;morphologicalcharacteristicsareexaminedand comparedwithterrestrialcounterparts,andimplications aboutformationmechanismsandexplorationpossibilities arediscussed. CavesareimportanttothefutureofMarsexploration becausetheyarebelievedtoprovideshelterfromarangeof harshsurfaceconditions,maintainingnear-pristinesurfacesandrelativelystablemicroclimates.Mars’sthinatmosphereandnegligiblemagneticfielddonoteffectively absorb,deflect,ormoderatenumeroushazards,including micrometeoroidimpacts,duststorms,extremetemperature variations,andhighfluxesofUV,alphaparticles,and cosmicrays(e.g.,Mazuretal.,1978;DeAngelesetal., 2002;Bostonetal.,2004;Cushingetal.,2007).Because organicmaterialscannotcontinuouslywithstandsuchhazards,cavesmaybeamongthefewhuman-accessiblelocationsthatpreserveevidenceofwhethermicrobiallife everexistedonMars.Cavesmayalsobecomevaluable resourcesforhumanexplorers,whowouldotherwisehave totransporttheirownsheltersorconstructtheminplace (e.g.,Horz,1985;CoombsandHawke,1992;Bostonetal., 2003).Additionally,exploringandcharacterizingMars’s volcaniccavesshouldenableustoconstraintheoriesabout lava-flowthermodynamicsandhydrodynamicsunderMars’s gravityandatmosphericconditions.Volcanicandothertypes ofcavesmayalsoprotectmineralformationsthateitherdo notformorbecomeburiedoralteredundersurfaceconditions(HillandForti,1997). WesurveyedallreleasedimagesfromMarsOdyssey’s ThermalEmissionImagingSystem(THEMIS)andMars ReconnaissanceOrbiter’sContextCamera(CTX)covering theflankflowsandlavaplainstothenorthofArsiaMons (235 u to243 u E, 2 8 u to2 u N,Fig.1)becausethisiswhere thefirstcave-entrancecandidateswereidentified(Cushing andTitus,2010).Entrancesintotwodistinctlydifferent cavetypesexistinthisregion:near-surfacelava-tubecaves thatformassingular,horizontal,curvilineartunnelsand volcano-tectoniccavesthatmayextenddownwardinto subsurfacefractures.Althoughwecannotdeterminehow faranyoftheMartiancavescontinuebeneaththesurface, itislikelythatthatmanycouldbeextensiveifourcomparisonswithterrestrialexamplesareappropriate,because terrestrialvolcaniccavescanbetensofkilometersinlength andhundredsofmetersdeep.Numeroussmallercave entrancesappeartoexistthroughoutthesampleregion, butthesecasesarenotdiscussedherebecausetheyaretoo G.E.Cushing–CandidatecaveentrancesonMars. JournalofCaveandKarstStudies, v.74,no.1,p.33–47.DOI:10.4311/ 2010EX0167R JournalofCaveandKarstStudies, April2012 N 33

PAGE 36

Figure1.ContextcameraimageshowingtheriftzoneandlavaplainsnorthofArsiaMonswiththedistributionofskylightbearingtube-fedlavaflows(A–H)andthevolcano-tectonicfracturesystem(I)listedinTable1(Image:MOLA128pixel-perdegreeshadedrelieffromJMARS(Christensenetal.,2007)). C ANDIDATECAVEENTRANCESON M ARS 34 N JournalofCaveandKarstStudies, April2012

PAGE 37

smallandpoorlyresolvedtobepositivelyidentifiedatthis time. S OURCE D ATA THEMISconsistsoftwocamerasthatrecordvisible andthermal-infraredwavelengths.Thevisible-wavelength cameraobservestheMartiansurfaceat18or36m/pixel infivediscretebandsthatcoverthevisibletonear-IR spectrumbetween 0.4and 0.9 m m(Christensenetal., 2004).Inthisstudy,weuseonlyband-3data( 0.654 6 0.053 m m)becauseitisacquiredinmostobservationsand providesthehighestsignal-to-noiseratio.At 100m/pixel, THEMISthermal-infrareddataaretoocoarsetodetect thefeaturespresentedhere.CTXdetectsabroadbandof visiblewavelengths( 0.5–0.8 m m)at 6m/pixel(Malin etal.,2007),allowingmoreprecisemeasurementstobe takenandenablingfeaturesthataresmallorambiguousin THEMISdatatobemoreclearlyidentified.AsofJune 2010,CTXcoveragedidnotincludemuchoftheregion northofArsiaMons,whereTHEMISidentifiedmanyof thetube-fedflowsandfracturesdiscussedhere.THEMIS Figure2.Comparisonofinflatedtube-fedlavaflows(andtheirassociatedsurfaceexpressions).Topimageshowsanunnamed flowinK laueavolcano’seasternriftzone.BottomimageshowsaproposedMartiantube-fedflowintheriftapronnorthof ArsiaMons(CTX:B02_010398_1751).Arrowsindicatelocationsofcandidateskylightentrances. G.E.C USHING JournalofCaveandKarstStudies, April2012 N 35

PAGE 38

andCTXdatasetsareappropriatetoconductsurveytypesearchesforcaveentrancesbecauseeachobservation covershundredsofsquarekilometes,andtheycumulativelycoverthemajorityofMars’stotalsurfacearea.The MarsReconnaissanceOrbiter’sHigh-ResolutionImaging ScienceExperiment(HiRISE)isavisible-wavelengthcamerathatobservesatultra-fineresolutiondownto 0.25m/ pixel,givingpixelsabout1/5000thesizeofaTHEMIS pixel(McEwenetal.,2007)thatallowsprecisemeasurementstobetakenandfine-scalemorphologiesandsurfacetexturestobeevaluated.BecausetheMROperforms routinerollmaneuversaboutitsdirectionofmotion, HiRISEandCTXcanobserveobjectsfromanoff-nadir perspective.Thisisusefulindeterminingwhetheravertical walloranoverhangingrimispresentaroundtheskylights. BecauseHiRISEwillcoveronlyasmallportionofMars duringitslifetime( 1%duringitsinitialtwo-yearprimary sciencephase(McEwenetal.,2007),itstargetsare carefullychosenandlimitedtospecificandscientifically relevantlocations,usuallybasedonTHEMISorCTX observations. L AVA -T UBE C AVES Lavatubesareinsulatedsubsurfaceconduitsthatallow, oronceallowed,flowinglavatomaintainhightemperatures,andthuslowviscosity,asittravelsfromasource venttodistantflowmargins(e.g.,WentworthandMacDonald,1953;Greeley,1971a,1971b,1972;Keszthelyi, 1995;Sakimotoetal.,1997).Thesearecommonstructures inbasalticpa hoehoe–typelavaflowsandfacilitatethe spreadofevolvedmaficlavaflowsbyenablinglarge volumestobetransportedoverlongdistances(Wentworth andMacDonald,1953;KeszthelyiandSelf,1998).Ifthe flowoflavathroughanestablishedtubebecomeseither diminishedordiverted,thenthetubemaydrainandleave anempty,intact,andstructurallycompetenttunnel.Surfaceentrances(skylights)mayexistwhereportionsoflavatubeceilingseithernevercompletelyformedorhave collapsed.Occurringfrequentlyinterrestrialbasalticvolcanism,lavatubesareexpectedtobecommoninMars’s volcanicregionsaswell(e.g.,Horz,1985;Keszthelyi,1995; Sakimotoetal.,1997). OnMars,numeroussinuousandlineardepressionsin volcanicregionsappearsimilartoorbitalviewsofinflated tube-fedlavaflowsonEarth(Fig.2).Thesedepressionsare channel-likeinappearanceandfollowsingle,sinuouspaths thatextendalongthecrestsof10to20mtopographicrises withlaterallyspreadingflanksthatoftenformchainsof eithertumuliorventstructures.Thismorphologyisa telltaleindicatorofatube-fedsystememplacedbyinflation,whichcanbeadominantpa hoehoemechanism acrosslowslopes(Kauahikaua,etal.,2003).Inflatedtubefedlavaflowsoftendevelopdownward-propagatingcracks (runningalongthedirectionofflow,Fig.2)thatform whenthecooledandhardenedlava-tubeceilingisraised duetointernalpressurefromsubsequentmagma-injection episodes(e.g.,Selfetal.,1998;Kauahikaua,etal.,2003, Glazeetal.,2005).However,someofthetubespresented herehavesectionswheretheaxialtrenchdoesnotsitatop atopographicrise,butinsteadcutsdownwardintothe surfaceandshowsnoevidenceofinflation(e.g.,Fig.3). Theseapparentlynon-inflatedlava-tubesectionsmayhave initiallybeganaschannelizedflowsthateventuallycrusted overtoforminsulatedconduits(e.g.,Wentworthand MacDonald,1953;Greeley,1971a,1971b).Lavatubesthat neitherbeginasopenchannelsnorbecomeinflatedby internalpressuresmayexhibitlittleornovisiblesurface expression,andoftencanonlybeidentifiedbythepresence ofskylightentrances(e.g.,Dagaetal.,1988;Greeley,1972; CalvariandPinkerton,1999;Miyamotoetal.,2005). Althoughthesurfacegroovesdescribedherearesometimes thoughttobecompletelycollapsedlavatubes,itisimportanttoemphasizethatthesefeaturesaremorelikelyto beeitherdilationalfracturesfrominflationortheupper surfacesofformerchannelizedflows.Theydonotnecessarilyindicatethediameterofanunderlyingevacuated tube,thatinternalcollapsehasoccurred,oreventhatthe tubesystemeverdrainedtoformanemptytunnel.Herewe suggestthatthepresenceofskylightentrancesintothese trenchesindicatesthatdrainingdidindeedoccuratthose locationsandthatlavatubesarelikelytoremainatleast partiallyintactbeneaththesurface(Figs.3and4top). Keszthelyi(1995)andSakimotoetal.(1997)modeledthethermodynamicsoflava-tubeformationtodeterminehowtubelengthsanddiametersonvariousplanetary bodieswouldberelatedtopropertiessuchaslavaviscosity, effusionrates,regionalslopes,gravity,andatmosphere. Thetube-fedflowsonMarsobservedtocontainskylights showsurfaceexpressionsupto71kmlong,whichis considerablyshorterthanthemaximummodeledlengthsof upto1000km,sothesetubes,whosesourceventsarenot seeninthevicinity,couldextendmuchfartherupslope. Theproposedlava-tubeskylightsdiscussedhereoccur withinseveraltube-fedflows(Table1).Theseentrancesare lessthan60macross,atleast10to30mdeep,anddonot extendlaterallybeyondthechannelsortrenchesthatcontain them.Insomecaseswhereinflationeffectsarenotapparent, theskylightsoccurinrelativelyflatsurfaceswheretheyhave causedwind-streakpatternstoform(Fig.4,top)andappeartobeatleastpartiallyfilledwithin-blowndust. Tobeclassifiedasacandidatelava-tubeskylightinthis paper,afeaturemusthaveatleastonepixelthatisdarker thantheshadowscastbynearbysurfacefeaturessuchas craterrims,existwithinasinuoustrenchthatrunsaxially alongatube-fedlavaflow,besubstantiallydeeperthanits hosttrench,andnothavecharacteristicsnormallyassociatedwithimpactcraterssuchasraisedrimsorejectapatterns. Bleacheretal.(2007a)observedlava-tubeskylightsin THEMISimageV11326014,buttheseareunlikelycaveentrancecandidatesbecausetheyarenotasdarkasnearby shadowedregionsandarenodeeperthantheirhosttrench. C ANDIDATECAVEENTRANCESON M ARS 36 N JournalofCaveandKarstStudies, April2012

PAGE 39

V OLCANO -T ECTONIC C AVES Theothertypeoflinearsurfacedepressionobservedto containcaveentrancesislocatedatthedistaledgeofthe ArsiaMonsnorthwest-flankventflows,approximately midwaybetweenArsiaMonsandPavonisMonsand 200kmeastoftheaforementionedtube-fedflows(Iin Fig.1).Thisstructureshowsevidenceofbothvolcanicand tectonicmechanismsandisclearlydifferentfromthe trenchesassociatedwithtube-fedlavaflowsinseveral ways.Itissubstantiallywiderthanthefracturesfound alongthecrestsoftube-fedflowsandiscomposedof10to 20kmlinearorcurvilinear,ratherthansinuous,segments thatcutacrossnumerouslavaflowsandconjoinatsharp angles,usuallybetween95and110 u (Figs.5,6,and7). Collectively,theconnectedfracturesextendmorethan 100km.Thisstructuredoesnotcontaintumuli,outflow vents,ortopographichighpoints,butinsteadhasgrabenlikedepressionswithbroad,flatfloorsandnoapparent verticaloffsetbetweenopposingwalls.Aregionaldust mantleuptoseveralmetersthickmasksanyparallel normalfaultsthatcouldindicatewhetheragraben-type collapseoccurred.Thelengths,segmentation,location,and orientationofthesefractures,whicharebetweenArsia MonsandPavonisMonsandalignedwiththeTharsisridgevolcanicsystem,suggesttheymayhaveformed throughdeeptectonicprocessesassociatedwiththeTharsis regionaluplift(Banerdtetal.,1992;Phillipsetal.,1990). Dilationaltectonicfracturesofthismagnitudecould extenddownwardasfaras5kmbeneaththesurface (Ferrilletal.,2003),wheretheywerelikelyintrudedand widenedbymagma,thusinducingformationoftheobservedgrabensandcaveentrances.Aswiththelava-tube skylights,thevolcano-tectoniccaveentrancesarenowider thanthefracturesinwhichtheyformed.However,unlike lava-tubecaves,whichtendtobesinuous,remainrelatively neartheuppersurface,andfollowregionalslopes, volcano-tectoniccavescouldextenddeepintotheirhost fracturesandbranchintocomplicatedsubsurfacenetworks (Coons,2010). Visibleevidencethatlow-viscositymaterialsflowedout ofthisfracturenetworkcanbeseenintheformof occasionalrimleveesandasubstantialoutflowfeature withdistinctlyfluvial-likestreamlinedbraidswherealarge volumeoffluidlaterallybreachedthefracturewallsinthe down-slopedirection(Fig.7).Ifthisfluidwaslava,thenits viscositymusthavebeenexceptionallylowtoformsucha fluvialpattern.Jaegeretal.(2010)rigorouslyexamineda similarexampleinAthabascaVallis,wherefluvial-like featureswerecarvedbymaficorultra-maficfloodbasalts withviscositiespossiblyaslowas10Pa ? s.Ontheother hand,thestreamlinedislandsinthisoutflowfeaturemay haveformedfromanoutbreakofliquidwater,asthe Tharsisregioncontainsmultipleexamplesoffluvial-type Figure3.HiRISEESP_016767_1785(alongflowFinFigure1).Collapsedlava-tubesectionthatÂ’sbeenpartiallyfilledby windblowndust.Caveentrancemightbeaccessiblebyavehicle. G.E.C USHING JournalofCaveandKarstStudies, April2012 N 37

PAGE 40

channels(Mouginis-Mark,1990;Mouginis-Markand Christensen,2005),andwater-icecloudsformaboveArsia Monsnearlyeverydayoftheyear(NoeDobreaandBell, 2005;Bensonetal.,2006). AnappropriateterrestrialcomparisontothisMartian volcano-tectonicfracturesystemmaybetheGreatCrack (Fig.6,top)inK laueavolcanoÂ’ssouthwestriftzone.The GreatCrackisalinearseriesoffractures,grabens,andpitcraterchainsthatextendfromK laueaÂ’scalderatothe ocean.Thisstructureformedwhenupwellingmagma intrudedintoadike-inducedfractureduringthe1823 K laueaflowevent,thuswideningthefractureandforming adikethattransportedmaterialmorethan35kmfromthe source.Normalfaultingandstopingabovethisdikecaused aseriespitcratersandgrabenstoform(OkuboandMartel, 1998),alongwithanumberofdeepandextensivecaves, someofwhichhavebeenexploredtodepthsexceeding180m (Coons,2010).AlthoughtheGreatCrackislikelyanontectonicfeaturecausedbyforcedmagmaticintrusion,while theMartiancrackswerelikelywidenedwhenmagmaflowed intopre-existingtectonicfractures,thesemechanismsboth involveinteractionsbetweenfracturesandmagmaandmay havecreatedsimilaropportunitiesformagmatodrainand producecomparablesurfaceexpressionssuchaspitsand fracturecavesatthesub-kilometerscale. Figure8showstheonlycurrentlyreleasedHiRISE observationofavolcano-tectonicfractureskylightwhere anoverhanging,dust-mantledrimandaninteriormound thatiseitherdustordust-coveredareclearlyresolved. Specificsofthisskylightarediscussedinthemeasurements section. A TYPICAL P IT C RATERS Adifferenttypeofcaveentranceisdescribedby Cushingetal.(2007),whodiscussanumberofsteepwalleddepressionsthatappeartobespecialcasesof commonpitcraters(Fig.4,bottom).AtypicalPitCraters (APC)arenotassociatedwithsurfacegroovesandare nearlyalwayscircular,withdiametersof 80to300m. HiRISEshowsthatsomeAPCsaredeep,sheer-walled cylindricalstructureswithnoapparentsubsurfaceextent, whileothersextendlaterallybeneathoverhangingrimsfor unknowndistances(Cushingetal.,2008).Figure9shows anAPCwhereacandidatecaveentrancecanbeseeninits floor;theimagehasbeencontrast-enhancedtodisplaythe instrumentÂ’slow-radiancelimit.Figure10showswhat appearstobeacommonpitcraterwithacentralskylight entranceintowhichsurfacematerialslikelycontinueto drain.ThisexampleispossiblyuniqueonMars,andit Figure4.Threedifferentcaveentrancecandidatetypeson Mars.Topimageshowsalava-tuberillewithmultiple skylightentrances(CTX:P17_007774_1757);thecenter imageisavolcano-tectonicfracturewithaskylightentrance (HiRISE:ESP_014380_1775,detailinFigure8);andthe bottomimageisanatypicalpitcraterwithadiameterof r 165m,adepth 245m,andanoverhangingrimof unknownextent(HiRISE:PSP_003647_1745). C ANDIDATECAVEENTRANCESON M ARS 38 N JournalofCaveandKarstStudies, April2012

PAGE 41

couldrepresentanintermediate,possiblyongoingformationstageofeitheranAPC,ifceilingcollapseprogresses outward,oracommonpitcrater,ifdrainingsurfacematerials eventuallyfilltheskylight.Haruyamaetal.(2009)recently discoveredalunarfeatureinanancientlavachannelthat looksidenticaltoMartianAPCsandsuggestitmaybea skylightentranceintoadeeplavatube.Table2showsa comparisonofthethreedifferentcave-entrancetypes. M EASUREMENTS TheMarsOrbiterLaserAltimeterontheMarsGlobal SurveyormeasuredsurfaceelevationsacrossMars(Zuber etal.,1992)withaverticalresolutionof 1m,andweused adatasetbinnedtoaspatialresolution(equatorial)of approximately256pixelsperdegree,or 230m.Weused theJavaMission-planningandAnalysisforRemote Sensing(JMARS)tool(Christensenetal.,2007)toregister MOLAwithvisibleobservationsanddeterminetheelevationprofileforeachfeaturecontainingatleastone candidateskylightentrance.Alltheobservedtube-fed flowsruncontinuouslydown-slopeatshallowanglesaveraging 0.5 u orless,andthevolcano-tectonicfractures average 0.3 u down-slopewithoccasionalminorrises (Fig.11;Table1). InTHEMISobservations,mostoftheskylightentrances areonthescaleofonlyafewpixelsorless,soaccurate dimensionswithreasonableerrorestimatescannotbe determinedfromthespatialextentofpixelsalone.Horizontaldimensionsoftheskylightswerecalculatedbyfittingan ellipsetothedarkpixelscorrespondingtoeachfeature, assumingthatpixelsoutsideoftheellipsematchthemedian pixelvalueofthescenewhilethoseinsideoftheellipsematch eithertheminimumpixelinthesceneorzeroDN(Digital Numbervalue).Thescenewasreconstructedassumingsubpixelmixingbetweentheinsideandoutsideoftheellipse, andtheestimatedDNforeachpixelwasweightedbythe areaoftheellipsecontainedwithinthatpixel.Thelocation, majorandminoraxes,androtationofthemajoraxiswere varieduntiltheroot-mean-squaredifferencebetweenthe actualscene(atthemedian-valuethreshold)andthe reconstructedscenewasminimized(Fig.12). Toestimatethedepth( d )ofafloorbeneathaskylight entranceinTHEMISimages,weassumedanadirviewing perspectiveanddividedthelength( D s )oftheinterior shadowcastbytheriminthedirectionofilluminationby thetangentoftheobservedsolarincidenceangle( i ).In mostcases,thelava-tubefloorsaresufficientlydeep, i is sufficientlylarge,andthefulldiameteralongthelineof illumination( D i )issufficientlysmallthattheshadowed/ sunlitboundarycannotbeseen(so D s 5 D i ),andonlya minimumvaluefordepthcanbeobtained.Assumingthat lava-tubefloorsmaintainessentiallyconstantdepths beneaththesurface,thelargestskylightsindicatethat interiorfloorsrunatleast15to30mbelowthesurface. Table1.Physicalcharacteristicsofskylight-hostinglavatubes(A-H)andvolcano-tectonicfractures(I)showninFigure1.THEMISIDsbeginwitha V,CTX IDsbeginwithaPoraB. Feature Primary ImageID a ImageScale (m/pixel) Incidence Angle(deg) Numberof Skylights Minimum Depth(m) Longitude Range(E) Latitude Range(N) TotalLength (km) a Elevation Range(m) Ave.Slope (deg) (A)V1261100817.979.904 10235.9to236.40.91to0.92 354164to42840.12 (B)V1843901317.766.724 18236.0to236.40.51to0.6332.54114to42470.23 (C)V1843901317.766.729 24235.8to236.80.05to0.4471.04020to43350.25 V1261100817.979.90 (D)V0527201817.776.605 10236.5to236.8 2 0.47to0.43 194293to43470.46 (E)V1198800117.678.894 12236.3to236.5 2 1.46to 2 1.38 154258to43390.31 (F)V2783700135.379.6032 23235.9to236.4 2 3.59to 2 2.8647.04457to47390.34 P20_008763_17495.357.21 B05_011532_17505.357.27 (G)B02_010398_17515.2857.895 19236.7to237.5 2 3.00to 2 2.77 554655to51760.54 B01_010187_17595.3257.77 (H)P09_004346_17575.3546.801 15237.8to238.3 2 2.73to 2 2.44 355086to53580.45 (I)V1646800117.670.419 35241.7to242.6 2 3.19to 2 2.24 1007480to69900.28 P17_007774_17575.3949.74 a Calculatedalongtheentireobservedlength. G.E.C USHING JournalofCaveandKarstStudies, April2012 N 39

PAGE 42

HiRISEhasobservedonefracture-caveskylightthus far,lookingslightlyoffnadiratanangleof6.8 u (see Fig.8).Thisimageshowsalargecentralmound,not apparentinTHEMISobservations,thatappearstobe composedofeitherdustordust-coveredrubble.Themajor andminoraxesofthisfeatureare 68mand 48m, respectively,whicharewithintheerrorboundsofour ellipsefittedtotheTHEMISdata.Basedonthesolar incidenceof35 u ,wecandeterminethatthedepthatthe edgeoftheshadowisapproximately37m.Accordingly,if slopesofthecentralmoundareattheangleofrepose( 30 to40 u ),thenthepeakofthemoundisapproximately19to 25mbeneaththesurfacelevel,andthenon-illuminated areadirectlybelowthewesternrimisatleast49to55m deepifthemoundslopesconsistentlydownwardtothat point.Theentranceledgeappearstocompletelyencircle thepitandslopesinwardfromthesurfacesothatthedust/ rockinterfacecanbediscerned.Assumingthesematerials lieattheangleofrepose,theupperoverhangingledgeis approximately5.5mto8.5mthick,withtheuppermost3 or4mbeingcomposedofdustmantling2.5mto4.5mof bedrock. Figure5.Asmallportionofthepropos edvolcano-tectonicfracturenetwork, shownheretoillustrateitsfracturelikeappearancefromorbit(CTX:P17_007774_1757).Theentirenetworkspans 100km.Figure6showspartof thisfeaturecomparedwithaterrestrialanalog,Figure7showsevidencethatlow-viscosityfluidflowedthrough thenetwork,andFigure8detailsoneofitsskylightentrances.Northisupward;arrowsindicatecandidate skylightentrances. C ANDIDATECAVEENTRANCESON M ARS 40 N JournalofCaveandKarstStudies, April2012

PAGE 43

Figure7.Evidencethatlow-viscosityfluidflowedthroughpre-existingtectonicfractures.DetailoffeatureIinFigure1; attheright,fluidlaterallybreacheditscontainmentandflowedacrossthesurface,creatingafluvial-likepattern (THEMISV16468001). Figure6.Volcano-tectonicfracturesystemsonEarthandMars.TopimageshowsaportionoftheGreatCrackinK lauea volcanoÂ’ssouthwestriftzone.BottomimageshowsaportionofaproposedMartianvolcano-tectonicfracturenetworklocated northofArsiaMons(THEMISV16468001). G.E.C USHING JournalofCaveandKarstStudies, April2012 N 41

PAGE 44

Figure8.DetailofskylightinHiRISE:ESP_014380_1775(alsoshowninthebottomofFig.4).Solarincidenceangle = 35 6 viewingangle = 6.8 6 fromvertical.Diameter = 68m(measuredfrominteriorrimedges).Northisupward.Approximate depths:A–B < 4m(dustlayer);B–C < 3m(bedrockoverhang);A–D < 22m(depthtotopofmound);A–E < 37m(depthto C ANDIDATECAVEENTRANCESON M ARS 42 N JournalofCaveandKarstStudies, April2012

PAGE 45

S UMMARYAND D ISCUSSION Thecaveentrancesandtheirrespectivehoststructures appeartobeyoungerthantheirsurroundings,notcovered orcrosscutbyotherflowsorfractures,buttheirabsolute agescannotbedeterminedwithcertaintybecausetube-fed flowscanbemuchyoungerthanthesurfacestheycross. However,theentiresampleregionislateAmazonianinage ( 500Ma)basedoncratersize-frequencydistributions (PlesciaandSaunders,1979;NeukumandHiller,1981; Werner,2005),withsomeareaspossiblyasyoungas50Ma (Werner,2005). E XPLORATION Severaltechnologiesmustbeadvancedbeforerobotic explorerscanvisitMartiancaves.Precisionlanding techniquesmustbedevelopedtoplacespacecraftatsuch smalltargets,andmeansofenteringandexploringcave environmentswhilemaintainingcontactwiththesurface mustalsobedeveloped(e.g.,Bostonetal.,2003,2004: Le veilleandDatta,2010).Furthermore,internationally agreedplanetaryprotectionpoliciesforbidanyvisitsto candidateastrobiologysitesuntilmicrobialcontamination issuescanbeaddressed(COSPAR,2008).Regardlessof thechallengesahead,ouridentificationofviablecave targetsisanimportantsteptowarddevelopinganentire newcategoryofplanetaryexplorationtechnologies. H UMAN H ABITAT P OTENTIAL Fromterrestrialexamples,lava-tubecaves,especially thosewithlong,unbrokenextentsandflat,relatively smoothfloors,areparticularlysuitabletohosthuman habitatsonMars.Additionally,lavatubesusuallyhave sufficientlateralextenttoisolatetheirinteriorsfromnearentrancesurfacehazards.Lava-tubecaves,bytheirnature, canprotectinhabitantsfromallofthehazardsthat humanswouldencounteronthesurface.Duststorms andmicrometeoroidscannotreachcaveinteriors,temperaturevariationsareminimizedincaveenvironments(e.g., Bostonetal.,2001;Wynneetal.,2008),androof thicknessesofonly1to2mcaneffectivelyshieldagainst alltypesofincomingradiation(NCRP,2001;DeAngeles, 2002). Foreithersmallexplorationteamsorcommunitiesof futurecolonists,lavatubesmaybethesafestandmost economicaloptiontoprotecthumansonMars.Transportingorconstructingsheltersthatadequatelyprotecthuman visitorsonthesurfaceofMarsmaybeunrealisticinterms ofresourceconsumption,especiallysincefeaturessuchas lavatubesthatcanprovidemuchofthenecessaryprotectionmayalreadybepresent. W ATERICE Besidesprovidingsuitableshelter,cavesarepotential reservoirsofstableormetastablewater-icedeposits,which couldbeaninvaluableresource.Beinghighlyinsulated environments,cavesthatextendatdownwardanglesinto thesurfacewilltrapandholdcoldairtoformisolated microclimates(e.g.,Balch,1900;Halliday,1954,Ingham etal.,2008).Ifconditionsarefavorable,thencaveicecould eithercondenseandaccumulateatratesthatexceed sublimationorcouldremaintrappedthereafterforming inmorefavorabletimes.Williamsetal.(2010)rigorously modeledcave-icedepositionandstabilityforvarious r shadowedge);A–F < 38m(depthdirectlybeloweastrim);A–G < 40m(maximumvisibledepth);A–H < 52m(depthdirectly belowwestrimassumingthemoundslopescontinuouslydownwardattheangleofrepose.)Valuesforthisfigurewere calculatedassuminganangleofreposeof35 6 Figure9.HiRISEimageESP_016622_1660.Rightpanelshowssameimageasleft,butwithcontraststretchedtodisplaythe instrument’slow-endradiancethresholdtorevealacandidateentrancetoacavethatmayextendfromtheupperright.Thispit is 65macrossand 45mdeep. G.E.C USHING JournalofCaveandKarstStudies, April2012 N 43

PAGE 46

locationsonMars,concludingthatsubsurfacewater-ice canexistforlongperiodsacrossmuchoftheplanetand thatconditionsareparticularlyfavorableintheTharsis region,inwhichmanycavesarelocated.Thesecavesarein advantageouslocationsforicetoaccumulatebecause water-icecloudsareobservedtoformnearlyeverydayof theMartianyearaboveandaroundArsiaMons(Bensonet al.,2003,2006;Maltagliatietal.,2008).Coldtrappingand icecondensationiswelldocumentedfornumerous terrestrialcaves(e.g.,Balch1900;Halliday,1954;Ingham etal.,2008)andoccurseveninariddesertenvironments. Thereisalsostrongevidencethatthewesternflankof ArsiaMon,whichisadjacenttoourstudyarea,becomes glaciatedduringperiodsofhighobliquity(Headand Marchant,2003;Sheanetal.,2007),andthermal-diffusion modelsindicatethatasubsurfaceicetableshouldstill persistunderneathadesiccatedlayerofdust(Helbertetal., 2006).Ificeenteredthenearbycavesduringtheseperiods, thensomeofitmaystillbepreservedbyacombination ofinsulationbyin-blownsurfacedust,coldtrapping,and theabsenceofsolarradiation.Ifsuchancientdepositsof water-iceindeedstillexist,theymaypreserveatmosphericsamplesfromthepastthatcontainvaluableinformationaboutMarsÂ’sclimatehistory(e.g.,Dansgaardetal., 1993). A STROBIOLOGY NearlyallknowncavesonEarthhostcommunitiesof microorganismsandcavesareamongthemostlikelyplaces tofindevidenceofpastorpresentmicrobiallifeonMars. However,theexamplespresentedhereareunlikelyprospects forastrobiologyduetotheirlocation.TheTharsisMontes volcanicregionisamongtheyoungestsurfacesonMarsand hadnotyetformedwhenconditionswereconduciveforthe developmentoflifeasweunderstandit.Accordingly,older caves,perhapsformedbyaqueousprocesses,wherelife mighthaveretreatedundergroundastheMartiansurface becameincreasinglyinhospitablearemorelikelytocontain evidenceofpastorextantlife(Grinetal.,1998;Cabroletal., 2009).Nonetheless,allmannerofextraterrestrialcave discoveriesarevaluabletothefutureofastrobiology becausetheyencouragestudyandpreparationforfuture cave-specificmissions. F UTURE W ORK ContinuingHiRISEobservationsofcaveentranceswill provideimportantdetailssuchasaccuratemeasurements andfine-scalemorphologiesthatcannotberesolvedby THEMISorCTX.Besidesallowingforhighlyaccurate dimensionalmeasurements,HiRISEmayalsoallowusto constrainvaluesforlocaldust-mantlethicknesses,nearentranceroofthicknesses(inoff-nadirobservations),and floorcharacteristics.Especiallyinlava-tubecaves,characteristicsoffloormaterialsthatliedirectlybeneathskylight entrancescanindicatewhentheskylightformedinrelation totheflow.Forexample,rubblepileswouldindicatethat roofcollapseoccurredsubsequenttoflowactivity,while smoothfloorscouldsuggestthatthecollapsedmaterials fellintoviscouslavaandeithersankorweretransported away.Smooth-floorlavatubeswillbeimportanttargets forfuturecave-explorationbecausetheyarelikelysafer andsimplertoenter,navigate,andpossiblyeveninhabit, comparedwithrough-flooredlavatubes,volcano-tectonic caves,oratypicalpitcraters. Thestructuresdiscussedhereexistwithinalimitedarea, andotherexamplesarelikelytoexistinothervolcanic Figure10.HiRISEimageESP_023531_1840.Themainpit is 195macrossandthecentralskylightis 40macrossat thewidestpoint.Assumingtheinnerslopesrestattheangle ofrepose( 35 6 ),theskylightisapproximately50mbelow thesurface,andshadowmeasurementsindicateafurther dropof 25mtothecenterofthefloor.Thisfeaturemay representanintermediateformationstageforeitheratypical pitcratersorcommonvolcanicpitcraters. Table2.Identificationcriteriafordifferentcave-entrancetypesaroundArsiaMons. Cave-EntranceTypeVolcanic?Tectonic?AssociatedRille?VisibleinIR?Circular? Lava-tubeSkylightYNYNOccasionally FractureSkylightYYYNOccasionally AtypicalPitCraters??NYY C ANDIDATECAVEENTRANCESON M ARS 44 N JournalofCaveandKarstStudies, April2012

PAGE 47

regionsacrossMars.THEMISandCTXdatawillbeused toconductaplanet-widesurvey,concentratingmostlyon volcanicregionswherecomparablefeaturesaremostlikely tooccur.AdditionalcaveentrancecandidateswillbesuggestedasHiRISEtargets.Cavesdetectedinolderregions, whichmaybemoresuitableastrobiologytargets,willbe comparedwiththosepresentedheretorevealdetailsabout howthesestructuresevolveovertimeduetocollapse, erosion,orburial. Futurethermal-infraredobservationscanprovide importantadditionalinformationaboutthesecavesthat isnotavailableatvisiblewavelengths.Diurnalandannual temperaturevariationsinsideacavecanbeinfluencedby itsdepthandoverallsubsurfaceextent(e.g.,Inghametal., 2008;Wynneetal.,2008),andcaveinternalsurfacetemperaturestendtorepresentthemeanannualtemperatureat thesurface(Cropley,1965;PflitschandPiasecki,2003; Wynneetal.,2008)unlesscomplexheat-transportmechanisms,suchasseasonalventilationduetomultipleentrances,exist.Forexample,deepcavesthattrapcoldair shouldhavemeaninternalsurfacetemperaturesthatare lowerthanaverage,whileshort,shallowcavescanbeinfluencedbysolarinsolationattheirentrancesandhave meantemperaturesthatarehigherthanaverage.Thecaves presentedherearetoosmalltobedetectedin100m/pixel THEMISIRdata,sobeforetheirtemperaturevariations canbeanalyzed,thenextgenerationofthermal-infrared camera,whichshouldhavearesolutionof10m/pixelor better,mustbeplacedinorbit. BecausevolcanicregionsareextensiveacrossMars,the discoveryofcaveentranceswithinthemhaslongbeen anticipated.Atleastthreedifferentcave-formingmechanismsappeartohaveoperatedintheflowfieldnorthof ArsiaMons:lavatubes,volcano-tectonicfracturesand atypicalpitcraters.Specificexamplescannowbetargeted byHiRISEandfutureinstrumentstoascertaintheir suitabilityforexplorationandhumanhabitationandto determinewhatcapabilitiesmustbedevelopedtoreach them. A CKNOWLEDGEMENTS WethanktheTHEMISandHiRISEscienceteamsfor supportingthisworkandtargetingthecaveentrances describedhere.ChrisOkubo,TimothyTitus,andLazlo Keszthelyiprovidedinvaluablesuggestionsandadvice. R EFERENCES Balch,E.S.,1900,Glacie `resorFreezingCaverns:Philadelphia,Allen, Lane&Scott,337p. Banerdt,W.B.,Golombek,M.P.,andTanaka,K.L.,1992,Stressand tectonicsonMars, in Kieffer,H.H.,Jakosky,B.M.,Snyder,C.W.,and Matthews,H.,eds.,Mars:Tucson,UniversityofArizonaPress, p.249–297. Benson,J.L.,Bonev,B.P.,James,P.B.,Shan,K.J.,Cantor,B.A.,and Caplinger,M.A.,2003,Theseasonalbehaviorofwatericecloudsin theTharsisandVallesMarinerisregionsofMars:MarsOrbiter CameraObservations:Icarus,v.165,p.34–52.doi:10.1016/S00191035(03)00175-1. Benson,J.L.,James,P.B.,Cantor,B.A.,andRemigio,R.,2006, InterannualvariabilityofwatericecloudsovermajorMartian Figure11.Lava-tubeelevationprofilesfrom128pixelper degreegriddedMarsOrbiterLaserAltimeterdata.Top profilerunsacrossaninflatedtube-fedflow;rillesarenot resolvedatthisresolution.Lowerprofilerunsalongthe lengthofatube-fedflow,showingitscontinuousdownslopecharacter. Figure12.Best-fitellipsesforsizefortwodifferenthaze values,withsurfacepixelssettothemediandigitalvalueof thescene. G.E.C USHING JournalofCaveandKarstStudies, April2012 N 45

PAGE 48

volcanoesobservedbyMOC:Icarus,v.184,p.365–371.doi:10.1016/ j.icarus.2006.03.014. Bleacher,J.E.,Greeley,R.,Williams,D.A.,Cave,S.R.,andNeukum,G., 2007a,TrendsineffusivestyleattheTharsisMontes,Mars,and implicationsforthedevelopmentoftheTharsisprovince:Journalof GeophysicalResearch,v.112,E09005.doi:10.1029/2006JE002873. Bleacher,J.E.,Greeley,R.,Williams,D.A.,Werner,S.C.,Hauber,E.,and Neukum,G.,2007b,OlympusMons,Mars:Inferredchangesinlate AmazonianagedeffusiveactivityfromlavaflowmappingofMars ExpressHighResolutionStereoCameradata:JournalofGeophysical Research,v.112,E04003.doi:10.1029/2006JE002826. Boston,P.J.,Spilde,M.N.,Northup,D.E.,Melim,L.A.,Soroka,D.S., Kleina,L.G.,Lavoie,K.H.,Hose,L.D.,Mallory,L.M.,Dahm,C.N., Crossey,L.J.,andSchelble,R.T.,2001,Cavebiosignaturesuites: Microbes,mineralsandMars:AstrobiologyJournal,v.1,no.1, p.25–55.doi:10.1089/153110701750137413. Boston,P.J.,Frederick,R.D.,Welch,S.M.,Werker,J.,Meyer,T.R., Sprungman,B.,Hildreth-Werker,V.,Thompson,S.L.,andMurphy, D.L.,2003,Humanutilizationofsubsurfaceextraterrestrialenvironments:GravitationalandSpaceBiologyBulletin,v.16,no.2, p.121–131. Boston,P.J.,Frederick,R.D.,Welch,S.M.,Werker,J.,Meyer,T.R., Sprungman,B.,Hildreth-Werker,V.,andThompson,S.L.,2004, Extraterrestrialsubsurfacetechnologytestbed:Humanuseand scientificvalueofMartiancaves:AmericanInstituteofPhysics ConferenceProceedings,v.699,p.1007–1018.doi:10.1063/1.1649667. Cabrol,N.A.,Grin,E.A.,andWynne,J.J.,2009,Detectionofcavesand cave-bearinggeologyonMars:40 th LunarandPlanetaryScience Conference,Houston,Texas,AbstractNo.1040. Calvari,S.,andPinkerton,H.,1999,LavatubemorphologyonEtna andevidenceforlavaflowemplacementmechanisms:Journalof VolcanologyandGeothermalResearch,v.90,p.263–280.doi:10. 1016/S0377-0273(99)00024-4. Christensen,P.R.,Jakosky,B.M.,Kieffer,H.H.,Malin,M.C.,McSween, H.Y.,Nealson,K.,Mehall,G.L.,Silverman,S.H.,Ferry,S.,and Caplinger,M.,2004,TheThermalEmissionImagingSystem (THEMIS)fortheMars2001OdysseyMission:SpaceScience Reviews,v.110,p.85–130,doi:10.1023/B:SPAC.0000021008.16305. 94. Christensen,P.R.,Gorelick,N.,Anwar,S.,Dickenshied,S.,Edwards,C., andEngle,E.,2007,NewinsightsaboutMarsfromthecreationand analysisofMarsglobaldatasets:AmericanGeophysicalUnionFall Meeting,SanFrancisco,California,AbstractNo.P11E-01. Coombs,C.,andHawke,R.,1992,Asearchforintactlavatubesonthe Moon:Possiblelunarbasehabitats, in Mendell,W.W.,ed.,The SecondConferenceonLunarBasesandSpaceActivitiesofthe21 st Century:NASAConferencePublication3166,v.1,p.219–229. Coons,D.,2010,CavingbeneaththeGreatCrackofKilaueaVolcano: NSSNews,v.68,no.12,p.8–16. COSPAR(CommitteeonSpaceResearch,WorldSpaceCouncilPlanetary ProtectionPolicy),Amended2011,http://0-science.nasa.gov.iii-server. ualr.edu/media/medialibrary/2011/0 2/16/Rummel_-_COSPAR_Activities_ to_PPS.pdf. Cropley,J.B.,1965,Influenceofsurfaceconditionsontemperaturesin largecavesystems:BulletinoftheNationalSpeleologicalSociety, v.27,p.1–10. Cruikshank,D.P.,andWood,C.A.,1972,LunarrillesandHawaiian volcanicfeatures:PossibleAnalogues:TheMoon,v.3,no.4, p.412–447. Cushing,G.E.,Titus,T.N.,Wynne,J.J.,andChristensen,P.R.,2007, THEMISobservespossiblecaveskylightsonMars:Geophysical ResearchLetters,v.34,L17201,doi:10.1029/2007GL030709. Cushing,G.E.,Titus,T.N.,Jaeger,W.L.,Kezthelyi,L.P.,McEwen,A.S., andChristensen,P.R.,2008,Continuingstudyofanomalouspit cratersintheTharsisRegionofMars:Newobservationsfrom HiRISEandTHEMIS,39 th LunarandPlanetaryScienceConference, AbstractNo.2447. Cushing,G.E.,andTitus,T.N.,2010,CavesonMars:Candidatesitesfor astrobiologicalexploration:AstrobiologyScienceConference2010, AbstractNo.5414. Daga,A.W.,Towarnicki,M.A.,Wendel,M.A.,andWendel,W.,1988, Evolvingconceptsoflunararchitecture:Thepotentialofsubselene development:Abstractsofthe2 nd LunarBasesandSpaceActivitiesof the21 st CenturyConference,Lunar&PlanetaryInstitute,67p. Dansgaard,W.,Johnsen,S.J.,Clausen,H.B.,Dahl-Jensen,D.,Gundestrup,N.S.,Hammer,C.U.,Hvidberg,C.S.,Steffensen,J.P., Sveinbjo ¨rnsdottir,A.E.,Jouzel,J.,andG.Bond,1993,Evidencefor generalinstabilityofpastclimatefroma250-kyrice-corerecord: Nature,v.364,p.218–220.doi:10.1038/364218a0. DeAngeles,G.,Wilson,J.W.,Clowdsley,M.S.,Nealy,J.E.,Humes, D.H.,andClem,J.M.,2002,LunarLavaTubeRadiationSafety Analysis:JournalofRadiationResearch,v.43supplement,p.S41– S45.doi:10.1269/jrr.43.S41. Ferrill,D.A.,etal.,2003,InfluenceofgravityonthegeometryofMartian normalfaults:34 th LunarandPlanetaryScienceConference,Abstract. No.2050. Glaze,L.S.,Anderson,S.W.,Stofan,E.R.,Baloga,S.,andSmrekar,S.E., 2005,Statisticaldistributionoftumulionpahoehoeflowsurfaces: AnalysisofexamplesinHawaiiandIcelandandpotentialapplications tolavaflowsonMars:JournalofGeophysicalResearch,v.110, B08202.doi:10.1029/2004JB003564. Greeley,R.,1971a,Observationsofactivelyforminglavatubesand associatedstructures,Hawaii:ModernGeology,v.2,p.207–223. Greeley,R.,1971b,LavatubesandchannelsintheLunarMariusHills: Earth,Moon,andPlanets,v.3,no.3,p.289–314.doi:10.1007/ BF00561842. Greeley,R.,1972,Additionalobservationsofactivelyforminglavatubes andassociatedstructures,Hawaii:ModernGeology,v.3,p.157–160. Grin,E.A.,Cabrol,N.A.,andMcKay,C.P.,1998,CavesintheMartian regolithandtheirsignificanceforexobiologyexploration:29 th Lunar andPlanetaryScienceConference,AbstractNo.1012. Halliday,W.R.,1954,IcecavesoftheUnitedStates:TheAmericanCaver, BulletinSixteenoftheNationalSpeleologicalSociety,v.16,p.3–28. Haruyama,J.,Hioki,K.,Shirao,M.,Morota,T.,Hiesinger,H.,vander Bogert,C.H.,Miyamoto,H.,Iwasaki,A.,Yokota,Y.,Ohtake,M., Matsunaga,T.,Hara,S.,Nakanotani,S.,andPieters,C.M.,2009, PossiblelunarlavatubeskylightobservedbySELENEcameras: GeophysicalResearchLetters,v.36,L21206.doi:10.1029/2009 GL040635. Head,J.W.,andMarchant,D.R.,2003,Cold-basedmountainglacierson Mars:WesternArsiaMons:Geology,v.31,no.7,p.641–644. doi:10.1130/0091-7613(2003)031 0641:CMGOMW 2.0.CO;2. Helbert,J.,Head,J.,Marchant,D.,Shean,D.,andKreslavsky,M.,2006, FirstProspectingforIceintheFlankDepositatArsiaMons:37 th LunarandPlanetaryScienceConference,AbstractNo.1371. Hill,C.A.,andForti,P.,1997,CaveMineralsoftheWorld,second edition:Huntsville,Alabama,NationalSpeologicalSociety,463p. Horz,F.,1985,Lavatubes:Potentialsheltersforhabitats, in Mendell, W.W.,ed.,LunarBasesandSpaceActivitiesofthe21 st Century: Houston,LunarandPlanetaryInstitute,p.405–412. Ingham,K.L.,Northup,D.,andWelbourn,W.C.,2008,Climate modelingfortwolavatubecavesatElMalpaisNationalMonument, NewMexicoUSA, in Espinasa-Peren a,R.,andPint,J.,eds., ProceedingsoftheX,XI,andXIIInternationalSymposiaon Vulcanospeleology:Austin,Texas,AssociationforMexicanCave StudiesBulletin19,p.126–131. Jaeger,W.L.,Keszthelyi,L.P.,Skinner,J.A.,Milazzo,M.P.,McEwen, A.S.,Titus,T.N.,Rosiek,M.R.,Galuszka,D.M.,HJowington-Kraus, E.,Kirk,R.L.,andtheHiRISETeam,2010,Emplacementofthe youngestfloodlavaonMars:Ashort,turbulentstory:Icarus,v.205, p.230–243.doi:10.1016/j.icarus.2009.09.011. Kauahikaua,J.,Sherrod,D.R.,Cashman,K.V.,Heliker,C.,Hon,K., Mattox,T.N.,andJohnson,J.A.,2003,Hawaiianlava-flowdynamicsduringthePu‘u‘O ‘o -Ku paianahaeruption:Ataleoftwo decades,inHeliker,C.,Swanson,D.A.,Takahashi,T.J.,eds.,The Pu‘u‘O ‘o -Ku paianahaEruptionofK laueaVolcano,Hawai‘i:The First20Years:U.S.GeologicalSurveyProfessionalPaper1676, p.63–87. Keszthelyi,L.,1995,Apreliminarythermalbudgetforlavatubesonthe Earthandplanets:JournalofGeophysicalResearch,v.100,no.B10, p.20411–20420.doi:10.1029/95JB01965. Keszthelyi,L.,andSelf,S.,1998,Somephysicalrequirementsforthe emplacementoflongbasalticlavaflows:JournalofGeophysical Research,v.103,no.B11,p.27447–27464.doi:10.1029/98JB00606. Le veille ,R.J.,andDatta,S.,2010,LavatubecavesonMars—seeking signsofpastlife:AstrobiologyScienceConference2010:Evolutionof Life:SurvivingCatastrophesandExtremesonEarthandBeyond, AbstractNo.5344. C ANDIDATECAVEENTRANCESON M ARS 46 N JournalofCaveandKarstStudies, April2012

PAGE 49

Malin,M.C.,BellIII,J.F.,Cantor,B.A.,Caplinger,M.A.,Calvin,W.M., Clancy,R.T.,Edgett,K.S.,Edwards,L.,Haberle,R.M.,James,P.B., Lee,S.W.,Ravine,M.A.,Thomas,P.C.,andWolff,M.J.,2007, ContextcamerainvestigationonboardtheMarsReconnaissance Orbiter:JournalofGeophysicalResearch,v.112,no.E5,E05S04. doi:10.1029/2006JE002808. Maltagliati,L.,Titov,D.V.,Encrenaz,T.,Melchiorri,R.,Forget,F., Garcia-Comas,M.,Keller,H.,Langevin,Y.,andBibring,J.-P.,2008, ObservationsofatmosphericwatervaporabovetheTharsisvolcanoes onMarswiththeOMEGA/MEXimagingspectrometer:Icarus, v.194,p.53–64.doi:10.1016/j.icarus.2007.09.027. Mazur,P.,Barghoorn,E.S.,Halvorson,H.O.,Jukes,T.H.,Kaplan,I.R., andMargulis,L.,1978,BiologicalimplicationsoftheVikingmission toMars:SpaceScienceReviews,v.22,p.3–34. McEwen,A.S.,Eliason,E.M.,Bergstrom,J.W.,Bridges,N.T.,Hansen, C.J.,Delamere,W.A.,Grant,J.A.,Gulick,V.C.,Herkenhoff, K.E.,Keszthelyi,L.,Kirk,R.L.,Mellon,M.T.,Squyres,S.W.,Thomas, N.,andWeitz,C.M.,2007,MarsReconnaissanceOrbiter’s HighResolutionImagingScienceExperiment(HiRISE):Journal ofGeophysicalResearch,v.112,no.E5,E05S02.doi:10.1029/ 2005JE002605. Miyamoto,H.,Haruyama,J.,Kobayashi,T.,Suzuki,K.,Okada,T., Nishibori,T.,Showman,A.,Lor enz,R.,Mogi,K.,Crown,D.A., Rodriguez,J.A.P.,Rokugawa,S. ,Tokunaga,T.,andMasumoto,K., 2005,Mappingthestructureandd epthoflavatubesusingground penetratingradar:GeophysicalR esearchLetters,v.32,L21316. doi:10.1029/2005GL024159. Mouginis-Mark,P.J.,1990,RecentwaterreleaseintheTharsisRegion ofMars:Icarus,v.84,p.362–373.doi:10.1016/0019-1035(90)900 44-A. Mouginis-Mark,P.J.,andChristensen,P.R.,2005,NewObservations ofvolcanicfeaturesonMarsfromtheTHEMISinstrument:JournalofGeophysicalResearch,v.110,E08007.doi:10.1029/2005 JE002421. NCRP,2001,RadiationProtectionGuidanceforActivitiesinLow-Earth Orbit:NationalCouncilonRadiationProtectionandMeasurements, NCRPPublicationNo.N.132. Neukum,G.,andHiller,K.,1981,MartianAges:JournalofGeophysical Research,v.86,no.B4,p.3097–3121.doi:10.1029/JB086iB04p03097. NoeDobrea,E.Z.,andBellIII,J.F.,2005,TESspectroscopicidentificationofaregionofpersistentwatericeontheflanksofArsia MonsVolcano,Mars:JournalofGeophysicalResearch,v.110, E05002.doi:10.1029/2003JE002221. Oberbeck,V.R.,Quaide,W.L.,andGreeley,R.,1969,Ontheoriginof Lunarsinuousrilles:ModernGeology,v.1,p.75–80. Okubo,C.H.,andMartel,S.J.,1998,PitcraterformationonKilauea volcano,Hawaii:JournalofVolcanologyandGeothermalResearch, v.86,p.1–18.doi:10.1016/S0377-0273(98)00070-5. Pflitsch,A.,andPiasecki,J.,2003,Detectionofanairflowsystemin Niedzwiedzia(Bear)Cave,Kletno,Poland:JournalofCaveandKarst Studies,v.65,p.160–173. Phillips,R.J.,Sleep,N.H.,andBarendt,W.B.,1990,Permanentupliftin magmaticsystemswithapplicationtotheTharsisRegionofMars: JournalofGeophysicalResearch,v.95,no.B4,p.5089–5100. doi:10.1029/JB095iB04p05089. Plescia,J.B.,andSaunders,R.S.,1979,Stylesoffaultingandtectonicsof theTharsisregion:10 th LunarandPlanetaryScienceConference, Abstract,p.986–988. Sakimoto,S.E.H.,Crisp,J.,andBaloga,S.M.,1997,Eruptionconstraints ontube-fedplanetarylavaflows:JournalofGeophysicalResearch, v.102,no.E3,p.6597–6613.doi:10.1029/97JE00069. Self,S.,Keszthelyi,L.P.,andThordarson,T.,1998,Theimportanceof pa hoehoe:AnnualReviewofEarthandPlanetarySciences,v.26, p.81–110. Shean,D.E.,Head,J.W.III,Fastook,J.L.,andMarchant,D.R.,2007, RecentglaciationathighelevationsonArsiaMons,Mars:Implicationsfortheformationandevolutionoflargetropicalmountain glaciers:JournalofGeophysicalResearch,v.112,E03004.doi: 10.1029/2006JE002761. Wentworth,C.K.,andMacdonald,G.A.,1953,StructuresandFormsof BasalticRocksinHawaii:WashingtonD.C.,GovernmentPrinting Office,U.S.GeologicalSurveyBulletin994,98p. Werner,S.C.,2005,MajoraspectsofthechronostratigraphyandgeologicevolutionaryhistoryofMars[Ph.D.dissertation]:Berlin,Frie University,http://www.diss.fu-berlin.de/2006/33/index.html,74p. Williams,K.E.,McKay,C.P.,Toon,O.B.,andHead,J.W.,2010,Doice cavesexistonMars?:Icarus,v.209,p.358–368.doi:10.1016/j.icarus. 2010.03.039. Wynne,J.J.,Titus,T.N.,andChongDiaz,G.,2008,Ondeveloping thermalcavedetectiontechniquesforEarth,theMoonandMars: EarthandPlanetaryScienceLetters,v.272,p.240–250.doi:10.1016/ j.epsl.2008.04.037. Zuber,M.T.,Smith,D.E.,Solomon,S.C.,Muhleman,D.O.,Head, J.W.,Garvin,J.B.,Alshire,J .B.,andBufton,J.L.,1992,The MarsObserverlaseraltimeterinvestigation:JournalofGeophysicalResearch,v.97,no.E5,p.7781–7797.doi:10.1029/92JE00 341. G.E.C USHING JournalofCaveandKarstStudies, April2012 N 47

PAGE 50

THEFIRSTSUBTERRANEANFRESHWATERPLANARIANS FROMNORTHAFRICA,WITHANANALYSISOF ADENODACTYLSTRUCTUREINTHEGENUS DENDROCOELUM (PLATYHELMINTHES, TRICLADIDA,DENDROCOELIDAE) A BDUL H ALIM H ARRATH 1,2 ,R ONALD S LUYS 3 ,A DNEN G HLALA 4 AND S ALEH A LWASEL 1 Abstract: Thepaperdescribesthefirstspeciesoffreshwaterplanarianscollectedfrom subterraneanlocalitiesinnorthernAfrica,representedbythreenewspeciesof Dendrocoelum O ¨ rsted,1844fromTunisiansprings.Eachofthenewspeciespossessesa well-developedadenodactyl,resemblingsimilarstructuresinotherspeciesof Dendrocoelum ,notablythosefromsoutheasternEurope.Comparativestudiesrevealed previouslyunreporteddetailsandvariabilityintheanatomyofthesestructures, particularlyinthecompositionofthemusculature.Anaccountofthisvariabilityis provided,anditisarguedthattheanatomicalstructureofadenodactylsmayprovide usefultaxonomicinformation. I NTRODUCTION TheFrenchzoologistsC.AlluaudandR.Jeannelwere amongthefirstworkerstoresearchinsomedetailthe subterraneanfaunaofAfrica(see,JeannelandRacovitza, 1914).Subsequently,anincreasingnumberofgroundwater specieswerereportedfromAfricancaves(Messana,2004). Althoughthecontinenthasnotbeenextensivelyexplored fromabiospeleologicalperspective,thelistofAfrican subterranean-dwellinganimalsislong(Messana,2004),but hasnotincludedflatworms.Thesubterraneanaquatic faunaofTunisiahasreceivedscantattention,despitethe presenceofalargenumberofwells,springs,andcaves, notablyinthenorthwesternpartofthecountry.Thefirst studyinTunisiaonthesubterraneanfaunaresultedinthe firstdescriptionofanewspeciesofThermosbaenacea, Thermosbaenamirabilis Monod,1924fromAfrica(Seurat 1921,1934).Duringthepastdecade,onlyafewexplorationsofsubterraneanhabitatsinTunisiahavebeen undertaken,resultinginthefindingofnewrecordsand newspeciesofgastropodsandcrustaceans(Juberthieetal., 2001;Ghlalaetal.,2009).However,subterraneanflatwormshavenotbeenpreviouslyreportedfromNorth Africa.Inthepresentpaper,wedescribethreenewspecies ofdendrocoelidfreshwaterflatwormsfromTunisia, representingthefirstplanarianstobereportedfrom subterraneanlocalitiesinNorthAfrica. ThefreshwaterplanarianfamilyDendrocoelidaebasicallyhasaHolarcticdistribution,albeitthatthereare majorareaswithinthisbiogeographicregionfromwhich specimenshavenotyetbeenreported(BallandReynoldson,1981,Fig.7).TheDendrocoelidaeareespecially diverseinthelakesOhridandBaikalandintheareaof theCarpathianMountains.FromLakeBaikal,especially, numerousmorphologicallycomplexgeneraandspecies havebeenreported(Porfirjeva,1977).TheHolarcticrange oftheDendrocoelidaeincludesthenorthwesternsectionof NorthAfrica,basedontherecordsof Dendrocoelum vaillanti DeBeauchamp,1955fromtheGrandeKabylie MountainsinAlgeriaand Acromyadeniummoroccanum De Beauchamp,1931fromBekritintheAtlasMountainsof Morocco(Sluys,2007,andreferencestherein).Bothspecies werecollectedfromepigeanhabitats,incontrasttothe speciesdescribedinthepresentpaper.Inadditiontothese recordsof D.vaillanti and A.moroccanum ,DeBeauchamp (1954)mentionedfindingtwoimmature,eyelessdendrocoelidsintheGrandeAtlasMountainsinMoroccoandan equallyimmature,ocellatedspecimeninA ¨ nDrahamin Tunisia. Allthreeofthenewspeciesdescribedinthepresent paperpossessaparticularcone-shapedstructureintheir copulatoryapparatusthatisalsopresentinmanyother dendrocoelids,amusculo-glandularorganoradenodactyl. Ourcomparativestudiesonotherspeciesofthegenus Dendrocoelum O ¨ rsted,1844revealedpreviouslyunreported detailsandvariabilityintheanatomyofthesestructures, particularlyinthecompositionofthemusculature.Inthis paper,weprovideafirstaccountoftheseanatomical detailsandtheirvariabilityandarguethattheanatomical structureofadenodactylsmayprovideusefultaxonomic information. *Correspondingauthor:halim.harrath@laposte.net 1 ZoologyDepartment,CollegeofScience,KingSaudUniversity,POBox2455, Riyadh11451,SaudiArabia 2 ResearchUnitAnimalReproductionandDevelopmentalBiology,Departmentof Biology,FacultyofSciencesofTunis,2092ManarII,Tunisia 3 InstituteforBiodiversityandEcosystemDynamics&NetherlandsCentrefor BiodiversityNaturalis,UniversityofAmsterdam(sectionZMA),P.O.Box94766, 1090GTAmsterdam,TheNetherlands 4 Unite deRecherchedeBiologieAnimaleetSyste matiqueEvolutive,Faculte des SciencesdeTunis,2092ManarII,Tunisia A.H.Harrath,R.Sluys,A.Ghlala,andS.Alwasel–ThefirstsubterraneanfreshwaterplanariansfromNorthAfrica,withananalysisof adenodactylstructureinthegenus Dendrocoelum (Platyhelminthes,Tricladida,Dendrocoelidae). JournalofCaveandKarstStudies, v.74,no.1,p.48–57.DOI:10.4311/2011LSC0215 48 N JournalofCaveandKarstStudies, April2012

PAGE 51

M ATERIALAND M ETHODS Flatwormswerecollectedfromthreedifferentnatural springsinTunisia.Eachspringfeedsanartificialreservoirof about1m 3 fromwhichwaterflowsthroughasmallchannel. Whenthereservoiriscompletelyclosedbyanirondooronits top,itisnotaccessible,andpenetrationofdaylightis obstructed.Subterraneanspeciesrandomlymigratetothe reservoir,where,apparently,t heyfindafavorableenvironment tolive,becausethecurrentinthechannelkeepssurfacespecies away.Specimenswerecollectedbysweepinganetseveraltimes throughthereservoir.Iftheirondoorwasclosed,andthusthe reservoirwasnotaccessible,theoutflowofthechannelwas blockedforatleast5minuteswithapieceofcloth.After removalofthecloth,theflood pulsewaschanneledthrougha nettocapturetheflatwormsp ecimens.Sampledspecimens weretransportedtothelabor atory,wheretheywerephotographed,fixedinSteinmann’so rBouin’sfluid,andpreserved in70%alcohol.Histologicalsectionsweremadeatintervalsof 7 m mandstainedinMallory-Cason.Drawingsofthecopulatoryapparatuswerefirstmadewiththehelpofacamera lucidaattachedtoacompoundmi croscope,thendigitized, and,subsequently,finalizedwithAdobeIllustratorCSand SnagIt.ThematerialexaminedisdepositedintheZoological MuseumoftheUniversityofAmsterdam(ZMA). S YSTEMATIC A CCOUNT Order Tricladida Lang,1884 Family Dendrocoelidae Hallez,1892 Genus Dendrocoelum O ¨ rsted,1844 Dendrocoelumconstrictum HarrathandSluys,sp.nov. (Figs.2,6,and7) Materialexamined. Holotype:ZMAV.Pl.6884.1,Ain SobahspringlocatedinthenorthwestofTunisia,onthe routelinkingTabarkatoTunis,approximately20kmfrom thesea,December2008,coll.H.Harrath&A.Ghlala, sagittalsectionson10slides. Paratypes:V.Pl.6884.2,ibid.,sagittalsectionson14 slides,V.Pl.6884.3,sagittalsectionsonelevenslides,V.Pl. 6884.4,horizontalsectionson6slides. Etymology. ThespecificepithetisderivedfromtheLatin adjective constrictus andalludestotheconstrictionor diaphragmthroughwhichtheseminalvesiclecommunicateswiththepeniallumen. Habitat. SpecimenswerecollectedfromAinSobahspring (36 u 87 9 20 0 N,8 u 55 9 01 0 E),situatedinawetzonein northwesternTunisia.Thislocalityisclosetotheroad fromTunistoTabarkaandlocatedatabout10kmfrom thelatter(Fig.1,localityA).Thewormswerefrequently foundattachedtogreenalgae;associatedfaunaconsisted ofisopodsandgastropods. Diagnosis. D.constrictum ischaracterizedbyanunpigmentedbody;presenceofrathersmalleyes;acommonvas deferensopeningintoaseminalvesicle,whichopensatthe tipofaconicalpapillaprojectingintotheanteriorsection Figure1.Samplinglocalities:A,AinSobah(Jendouba);B,AinDarForn(Siliana);C,AinElAjmi(Kef). A.H.H ARRATH ,R.S LUYS ,A.G HLALA,AND S.A LWASEL JournalofCaveandKarstStudies, April2012 N 49

PAGE 52

ofthespaciouspeniallumen,thusformingakindof diaphragm;principallydorsaltestes,butwithalsoafew ventralfollicles;anadenodactylonlyslightlysmallerthan thepenispapilla. Description. Maturespecimensinfullystretchedstateare upto12mmlongand2mmwide(Fig.2).Thebodyis principallyunpigmented,butappearsbeige,red,blackor slightlygreen,dependingontheintestinalcontents.The rectangularheadhasrathersmalleyesthatareplaced relativelyfarfromtheanteriorendofthespecimen. Althougheyesarealwayspresent,theyarenotalways clearlyvisibleinlivespecimens.Anadhesiveorganwasnot evidentinliveorpreservedspecimens.Histologicalsections revealedthepresenceofonlyaverypoorlydeveloped,noncupshapedadhesiveorgan. Theunpigmentedpharynxispositionedentirelyinthe posteriorhalfoftheanimal.Themuscularsystemofthe internalzoneisformedbyintermingledcircularand longitudinalfibers,whichconformswithcharacteristic pharynxofthefamilyDendrocoelidae(SluysandKawakatsu,2006,andreferencestherein). Thetestesaresituatedoneithersideofthebody;they areessentiallydorsalinposition,butwithafew,distinctly Figures2–5.2. Dendrocoelumconstrictum ,externalfeaturesofalivespecimen.3. Dendrocoelumduplum ,externalfeatures ofalivespecimen.4. Dendrocoelumamplum ,headofalivespecimen.5. D.amplum ,posteriorendofalivespecimen.See captiontoFigures6–11forkey. T HEFIRSTSUBTERRANEANFRESHWATERPLANARIANSFROM N ORTH A FRICA,WITHANANALYSISOFADENODACTYLSTRUCTUREINTHEGENUS D ENDROCOELUM (P LATYHELMINTHES ,T RICLADIDA ,D ENDROCOELIDAE ) 50 N JournalofCaveandKarstStudies, April2012

PAGE 53

ventralfollicles.Thetestesextendfromthelevelofthe ovariestoalmosttheposteriorendofthebody.The hemisphericalpenisbulbconsistsofintermingledlongitudinalandcircularmusclefibers(Fig.6).Themoreorless conicalpenispapillaiscoveredwithamoderatelythick, nucleatedepitheliumthatisunderlainwithalayerof circularmuscle,followedbyathinlayeroflongitudinal fibers.Thetwovasadeferentiarunventrally,butatthe levelofthecopulatorybursa,theycurvetothedorsalside andseparatelyenterthepenisbulb,afterwhichtheducts unitetoformacommonvasdeferensthatopensintoa seminalvesicle.Thelatteropensatthetipofthepapilla, whichprojectsintothevoluminousproximalsectionofthe peniallumen,thusformingakindofdiaphragm.Thiswide cavityislinedwithathick,nucleatedepithelium. Theadenodactyl(Fig.7)isslightlysmallerthanthe penisandsituatedtotherightofthelatter.Ithasa sphericalandverymuscularbulb,withintermingled longitudinalandcircularmusclefibersandanelongated papilla.Theadenodactylcontainsaratherlongtubular lumen.Throughouttheparenchymaofthepapillaextends alayeroffinecircularmusclefibersthatstainsbrightblue. Ectallytothiszoneofcircularmusclesrunsalayerof longitudinalmusclefibers.Entallytothebrightbluezone ofcircularmusclesthereisalsoathinlayeroflongitudinal fibers. Theovariesaresituatedataboutonethirdofthe distancebetweenthebrainandtherootofthepharynxand occupytheentiredorso-ventraldiameterofthebody.The oviductsarisefromthepostero-dorsalwalloftheovaries andrunventrallycaudadtothelevelofthecopulatory apparatus.Theyunitetoformacommonoviductjust behindthecopulatoryapparatus.Thiscommonoviduct receivesthesecretionofshellglandsandopensintothe posteriorpartofthecommonatrium. Thecopulatorybursaisalargesacthatoccupies almosttheentiredorsal-ventraldiameterofthebodyand issituateddirectlybehindthepharynx.Thebursais coveredbyathinlayeroflongitudinalmusclesandin maturespecimenshasalargecavitycontainingremnants ofaspermatophore.Thebursalcanalarisesfromthe postero-dorsalwallofthebursaandrunslatero-dorsally tothemaleatriumandthepenisbulb.Itislinedwitha thick,nucleatedepitheliumandissurroundedbytwo layersoflongitudinalmuscl es.Atitsdistalpart,just abovethecommonatrium,thebursalcanalwidens considerablyandiscovered withalayerofsubepidermal circularmuscles,followedbyalayeroflongitudinal fibers. Discussion. Thereareonlyfourotherspeciesof Dendrocoelum forwhichacommonvasdeferenshasbeen reported: D.puteale Kenk,1930, D.kenki DeBeauchamp, 1937, D.jablanicense (Stankovic andKoma rek,1927),and D.botosaneanui DelPapa,1965.Therefore,wewillrestrict ourcomparativediscussiontothesefourspecies.In contrastto D.constrictum ,thespermductsfuseoutside ofthepenisbulbin D.botosaneanui D.puteale ,and D. kenki (DeBeauchamp,1932,1937;DelPapa,1965).Itis onlyin D.jablanicense thatthespermductsfirstenterthe penisbulbandthenfusetoformacommonvasdeferens (Kenk,1978),asisthecasewith D.constrictum .However, in D.jablanicense thetestesareventral,contrastingwith thedorsallyplacedfolliclesin D.constrictum .In D. botosaneanui and D.puteale thetestesaresituatedcentrally anddorsallyinthebody,respectively,whereasin D.kenki theyhaveaventralposition. Apartfromthepointoffusionofthevasadeferentia andthepositionofthetestes,thereareafewotherdetails ofthecopulatoryapparatusworthmentioninginwhich D. constrictum eitherdiffersfromoragreeswiththefourother speciesof Dendrocoelum mentionedabove. D.constrictum resembles D.jablanicense D.kenki ,and D.puteale inthat theyalllackaflagelluminthepenispapilla,whereas D. botosaneanui doeshaveaflagellum.In D.constrictum D. jablanicense ,and D.kenki theadenodactylissituatedon therightsideofthemidlineofthebody,whereasin D. puteale itislocatedontheleftside;theprecisesituationin D.botosaneanui isnotclearfromitsdescription.Inboth D. constrictum and D.botosaneanui theadenodactylis somewhatsmallerthanthepenis,whereasin D.jablanicense and D.puteale theadenodactylisbiggerthanthe maleorgan;in D.kenki theadenodactylandthepenisare aboutthesamesize. Both D.constrictum and D.jablanicense possesseyes, whereastheseareabsentin D.botosaneanui D.kenki ,and D.puteale Acharacteristicfeatureof D.constrictum isthatthe seminalvesicleopensatthetipofaconicalpapilla projectingintotheproximal,anteriorsectionofthelumen ofthepenispapilla.Asimilarprojectionisalsopresentin D.kenki and D.racovitzai DeBeauchamp,1949,butonthe basisofotherfeatures D.constrictum cannotbeequated witheitherofthesetwospecies. Dendrocoelumduplum HarrathandSluys,sp.nov. (Figs.3,8,and9) Materialexamined. Holotype:ZMAV.Pl.6885.1,AinEl Ajmispring,locatedinthenorthwestofTunisia,at12km fromDahmanivillage,December2008,coll.H.Harrath andA.Ghlala,sagittalsectionson12slides. Paratypes:V.Pl.6885.2,ibid.,sagittalsectionson8 slides. Etymology. ThespecificepithetisderivedfromtheLatin adjective duplus ,double,andreferstothepresenceoftwo constrictionsordiaphragmsinthemaleductrunning throughthepenialpapilla. Habitat. ThematerialexaminedwascollectedfromAinEl Ajmispring(Fig.1,localityC)(35 u 51 9 03 0 N,8 u 47 9 52 0 E). A.H.H ARRATH ,R.S LUYS ,A.G HLALA,AND S.A LWASEL JournalofCaveandKarstStudies, April2012 N 51

PAGE 54

Figures6–11.6. Dendrocoelumconstrictum .ZMAV.Pl.6884.1,sagittalreconstructionofthecopulatoryapparatusatthe levelofthepenis(anteriortotheright).7. D.constrictum. ZMAV.Pl.6884.1,sagittalreconstructionofthecopulatory apparatusattheleveloftheadenodactyl(anteriortotheright).8. Dendrocoelumduplum. ZMAV.Pl.6885.1,sagittal reconstructionofthecopulatoryapparatusatthelevelofthepenis(anteriortotheright).9. D.duplum. ZMAV.Pl.6885.1, sagittalreconstructionofthecopulatoryapparatusattheleveloftheadenodactyl(anteriortotheright).10. Dendrocoelum amplum. ZMAV.Pl.6886.1,sagittalreconstructionofthecopulatoryapparatusatthelevelofthepenis(anteriortotheleft). 11. D.amplum. ZMAV.Pl.6886.1,sagittalreconstructionofthecopulatoryapparatusattheleveloftheadenodactyl(anterior T HEFIRSTSUBTERRANEANFRESHWATERPLANARIANSFROM N ORTH A FRICA,WITHANANALYSISOFADENODACTYLSTRUCTUREINTHEGENUS D ENDROCOELUM (P LATYHELMINTHES ,T RICLADIDA ,D ENDROCOELIDAE ) 52 N JournalofCaveandKarstStudies, April2012

PAGE 55

ThislocalityisnotveryfarfromtheRomanruinsofEl Medeinaandlocatedabout12kmfromthetownof Dahmani.Accompanyingfauna:alargenumberofisopods. Diagnosis. D.duplum ischaracterizedbyanunpigmented body;thepresenceofbroadauricularlobes;testesthatare locateddorsallyaswellasventrally;analmostvertically orientedpenispapilla;adiaphragminthedistalsectionof theejaculatoryduct;anejaculatoryductthatopensatthe tipofabluntpapilla,projectingintothepeniallumen,thus formingakindofseconddiaphragm;anadenodactyl papillathatisconsiderablysmallerthanthepenispapilla. Description. Livinganimalsupto14mmlongand1.4mm wide.Dorsalsurfacewhite.Headoffreelymovinganimals withatruncatedanteriormarginandadistinct,broadly roundedauricularlobeoneitherside;eyesabsent(Fig.3). Anadhesiveorganwasnotevidentinliveorpreserved specimens. Testesaresituateddorsally,aswellasventrally, extendingthroughoutthebodyfromdirectlybehindthe ovariestoalmostthetailend.Thereareonlyafew(3to4) testesintheprepharyngealpartofthebody,withthe majorityofthefolliclesbeinglocatedinthepostpharyngeal part.Thepenisisprovidedwitharelativelysmall hemisphericalmuscularbulbandanelongated,moreor lessverticallyorientedcylindricalpapilla(Fig.8).The papillaiscoveredwithaflat,nucleatedepitheliumthatis underlainbyalayerofcircularmusclefibersandavery thinlayeroflongitudinalmusclefibers.Thetwovasa deferentiarunventrally,andwhentheyarriveatthelevel ofthecopulatorybursa,theywidentoformtwoverylarge spermiducalvesicles.Afterhavingnarrowedandpenetratedthepenisbulb,thespermductsimmediatelyuniteto formacommonintrabulbarvasdeferensthatextendsas theejaculatoryduct(Fig.8).Throughadiaphragmthis ejaculatoryductcommunicateswithadistal,narrow sectionofthepenislumenthatopensatthetipofablunt papillaandprojectsintotheproximalpartofthelumenof thepenispapilla,formingakindaseconddiaphragm.This penislumenreceivestheabundantsecretionoferythrophilicpenisglands. Theadenodactyl(Fig.9),situatedventrallytotheleft ofthepenis,isalittlelargerthanthepenisandisprovided withaconicalpapillaandaspherical,well-developedbulb. Thisadenodactylisprovidedwithadistinctlylight-bluestainingzoneofcircularmusclesthatrunsthroughthe mesenchymefromthebulbtothetipofthepapilla.Ectally tothisbluezoneofcircularmusclesrunsalayerofredstaininglongitudinalmuscles.Athinlayeroferythrophilic longitudinalmusclesisalsopresententallytothebluezone offinecircularmusclefibers.Noglandswereobservedto dischargeintothelargelumenoftheadenodactyl. Thetwoovariesaresituatedjustabovetheventral nervecordaboutonequarterofthedistancebetweenthe brainandtherootofthepharynx.Theyarerelativelylarge andcanbeobservedinalivingmaturespecimenastwo whitishspots.Thetwooviductsunitetoformarelatively longcommonoviductthatopensintothecommonatrium. Thecopulatorybursaisasmallsac,situateddorsallyand directlybehindthepharyngealpouch.Fromthebursaruns thebursalcanal,whichnarrowsatitsproximalsectionand isonlysurroundedbyathinlayeroflongitudinalmuscles. Thereaftertheductwidens,andatitsvertically,dorsoventrallyrunningsection,itissurroundedbyathick, subepitheliallayerofcircularmuscles,followedbyathin layeroflongitudinalmuscles;thissectionofthecanal opensintothedorsalpartofthecommonatrium. Discussion. Mostspeciesof Dendrocoelum haveamoreor lessdevelopedauricularlobeoneithersideofthehead.But thebroadlobesof D.duplum havenotbeendescribed foranyotherspecies.Onlytheauricularlobesof D. adenodactylosum (Stankovic andKoma rek,1927)approachtheconditionof D.duplum (Kenk,1978,Fig.5). Withrespecttothecommunicationofthevasa deferentiawiththepenis,fourconditionscanbedistinguishedwithinthegenus Dendrocoelum :(1)Inthemajority ofthespecies,thevasadeferentiaseparatelypenetratethe penisbulbandsubsequentlyopenintoanintrabulbar seminalvesicle.(2)Inasmallernumberofspecies,the spermductsopenintothesomewhatexpandedproximal portionoftheejaculatoryduct.(3)Inanumberofspecies, thevasadeferentiaopenseparatelyintoamuchmoredistal sectionoftheejaculatoryduct.(4)Inonlyfivespecies,the vasadeferentiafusetoformanintra-(condition4a)or extrabulbar(condition4b)commonvasdeferens.Forthese conditions,seediscussionaboveon D.constrictum ;the differencebetweencondition(4a)andcondition(2)resides inthefactthatintheformertheproximalsectionof theejaculatoryductishardlyornotatallexpanded. Conditions(2)and(4a)alsoholdtruefor D.duplum r totheright).Abbreviationsusedinthefigures:ab:adenodactylbulb;ad:adenodactyl;ap:adenodactylpapilla;bc:bursalcanal; ca:commonatrium;cb:copulatorybursa;cm:circularmuscles;co:copulatoryapparatus;cod:commonoviduct;cvd:common vasdeferens;d:diaphragm;e:eye;ed:ejaculatoryduct;gl:glands;go:gonopore;h:head;ic:intestinalcaecum;l:lumen;lm: longitudinalmuscles;ma:maleatrium;o:ovary;od:oviduct;pb:penisbulb;pg:penisglands;ph:pharynx;pp:penispapilla;sg: shellglands;sp:spermatophore;spv:spermiducalvesicle;sv:seminalvesicle;t:tail;vd:vasdeferens;te:testes. A.H.H ARRATH ,R.S LUYS ,A.G HLALA,AND S.A LWASEL JournalofCaveandKarstStudies, April2012 N 53

PAGE 56

Amongthespeciesforwhichconditions(2)or(4a)are applicable,thegrossmorphologyofthecopulatory apparatusof D.duplum resemblesthatof D.banaticum CodreanuandBalcesco,1967, D.cavaticum (Fries,1874), D.chappuisi DeBeauchamp,1932, D.coiffaiti De Beauchamp,1956, D.maculatum (Stankovic andKoma rek, 1927), D.polymorphum CodreanuandBalcesco,1967, D. stenophallus CodreanuandBalcesco,1967, D.tuzetae Gourbault,1965,and D.vaillanti DeBeauchamp,1955. However,inallofthesespecies,thevasadeferentia separatelytraversethepenisaconsiderabledistancebefore communicatingwiththeejaculatoryduct,whereasin D. duplum, thespermductsjoinimmediatelyaftertheyhave penetratedthedorsalpartofthepenisbulb.Furthermore, noneofthesespeciesshowthedoublevalveordiaphragm systemof D.duplum. Eyesareabsentin D.duplum D.banaticum D. chappuisi D.polymorphum D.stenophallus ,and D.tuzetae whereastheyarepresentin D.cavaticum D.coiffaiti D. maculatum ,and D.vaillanti Dendrocoelumamplum HarrathandSluys,sp.nov. (Figs.4,5,10,and11) Materialexamined. Holotype:ZMAV.Pl.6886.1,AinDar Fornspring,nearBargouvillage,locatedinthegovernorateofSiliana,December2008,coll.H.HarrathandA. Ghlala,sagittalsectionson10slides. Etymology. ThespecificepithetisderivedfromtheLatin adjective amplus ,spacious,andalludestothedistinct wideningofthebursalcanalbeforeitopensintotheatrium. Habitat. AnimalswerecollectedfromDarFornspringin theBargouMountains(35 u 58 9 52.75 0 N,9 u 30 9 43.32 0 E),at approximately1kmnorthoftheroadfromSilianato Ousletia(Fig.1,localityB).Associatedfauna:chiefly isopods. Diagnosis. D.amplum ischaracterizedbyawhitebody;the presenceoftwosmalleyes;largetestes,situatedboth dorsallyandventrally;twointrabulbarseminalvesicles, eachreceivingtheopeningsofavasdeferens;acommon intrabulbarseminalvesiclethatopensviaadiaphragminto thespaciouspenialcavity;abursalcanalthatformsa considerableexpansionatthevaginalregion;anadenodactylthatissmallerthanthepenis. Description. Livingspecimenupto8mmlongand1.6mm wide.Dorsalsurfacedirtywhite.Headslightlyrounded, withtwosmalleyessituatedatequaldistancefromeach otherandfromthemarginsofthehead(Fig.4).The anteriorendoftheanimalisprovidedwithanadhesive organ,consistingofaconsiderableinvaginationthatis linedwithaninfranucleatedepidermis,penetratedby erythrophilicglands. Thenumerous,largetestesaresituateddorsally,aswell asventrally,beginningattheleveloftheovariesand extendingtoalmostthetailend(Fig.5).Thesubspherical penisbulbismuscularandhousestwolargeseminal vesiclesthataresituatedclosetoeachotherandfilledwith spermintheholotypespecimen(Fig.10).Eachseminal vesiclereceivesatitsanteriorsectiontheopeningofa spermductand,subsequently,opensintoasmaller commonvesicle.Thelatterisseparatedbyalarge diaphragmfromavoluminouspenialcavity.Thiswide cavityislinedwithathick,nucleatedepithelium.The papillaiscoveredwithathick,nucleatedepithelium,which isunderlainbyalayerofcircularmusclefibers. Theadenodactyl,locatedtotheleftofthepenis,is smallerthanthelatterandisprovidedwithawidelumen thatreceivesthesecretionofglands(Fig.11).Inthe muscularhemisphericalbulbthislumenexpandstoforma cavityofregularoutlinethatissurroundedbyintermingled layersoflongitudinalandcircularmusclefibers.The adenodactyllumenfollowsacentralcoursethroughthe conicalpapillaandopensterminally.Alayeroffine circularmusclefibersthatstainsbrightblueextends throughouttheparenchymaofthepapilla.Ectallyand entallytothiszoneofcircularmusclesrunsalayerof longitudinalmusclefibers. Thetwosphericalovaries,about0.22mmindiameter, aresituatedveryclosetothebrain(Fig.4).Theoviducts arisefromthepostero-ventralsideoftheovariesandrun ventrallytothelevelofthecopulatoryapparatus,where theycurvedorsallyandunitetoformacommonoviduct, whichopensintothefemaleatrium.Thecommonoviduct, aswellasthedistalsectionsoftheoviducts,receivesthe secretionofshellglands.Thebursalcanalislinedwith athickepitheliumandsurroundedbyathinlayerof longitudinalmusclefibers.Thecanalrunsdorsallytothe copulatoryapparatustocommunicatewithasac-shaped copulatorybursa,situatedclosetotheposteriorwallofthe pharyngealpouchandlinedwithatallepithelium.The bursalcanalexpandsconsiderablyindiameteratthe vaginalarea,whereitreceivesthesecretionofwelldevelopedshellglandsandsubsequentlyopensintothe femaleatriumviaaslightconstriction. Discussion. Severalspeciesof Dendrocoelum aresimilarto D.amplum inpossessingeyes,avasadeferentiaexpanded toformlarge,introbulbarseminalvesicles,orthepresence ofadistinctwideningofthebursalcanal.Theyaresimilar totheconditionsfoundin D.amplum arethefollowing: D. adenodactylosum (Stankovic andKoma rek,1927), D. maculatum (Stankovic andKoma rek,1927), D.jablanicense (Stankovic andKoma rek,1927), D.lacustre (Stankovic ,1938), D.plesiophthalmum DeBeauchamp,1937. Inboth D.adenodactylosum and D.maculatum the adenodactylisverylarge(Kenk,1978),ascomparedwith thesizeofthepenis,thuscontrastingwiththecondition in D.amplum ,inwhichtheadenodactylandpenisareof T HEFIRSTSUBTERRANEANFRESHWATERPLANARIANSFROM N ORTH A FRICA,WITHANANALYSISOFADENODACTYLSTRUCTUREINTHEGENUS D ENDROCOELUM (P LATYHELMINTHES ,T RICLADIDA ,D ENDROCOELIDAE ) 54 N JournalofCaveandKarstStudies, April2012

PAGE 57

aboutthesamelength. D.maculatum shareswith D. amplum thepresenceofbothdorsalandventraltestes, albeitthatthefolliclesarecomparativelysmallinthe formerandlargeinthelatter. In D.jablanicense, thebursalcanalmaycommunicate withthecommonatriumthroughoneortwoopenings (Kenk,1978).However,inthisspeciesthevasadeferentia fusetoformashort,intrabulbarcommonvasdeferens, whichopensintoaseminalvesicle.Thissituationis stronglydifferentfromthecourseofthevasadeferentia in D.amplum Although D.lacustre doespossessawidenedbursal canalnearthepointofcommunicationwiththeatrium,it lackstheexpanded,intrabulbarseminalvesiclesthatare characteristicfor D.amplum Judgingfromthedrawingofthecopulatoryapparatus publishedbyDeBeauchamp(1937,Fig.3), D.plesiophtalmum doespossessthewidenedsectionintheposterior partofthebursalcanal,butlacksthewidened,intrabulbar portionsofthevasadeferentia.Thelong,conicaland pointedpenispapillaof D.plesiophthalmum israther differentfromthestubbypapillaof D.amplum .Unfortunately,theprecisemorphologyandcourseoftheejaculatoryductthroughthepenispapillaisunknownfor D. plesiophthalmum A NATOMYOF A DENODACTYLSIN D ENDROCOELUM Thepresenceandpreciselocationofamusculoglandularorganoradenodactylinthecopulatorycomplex ofdendrocoelidplanarianshasgainedsometaxonomic importance.Forexample,thegenus Dendrocoelopsis Kenk, 1930ischaracterizedbythefact,amongothers,thatit lacksanadenodactyl,whileseveralsubgeneraof Dendrocoelum havebeenproposedbasedpartlyontheshape,size, andlocationoftheadenodactyl(Gourbault,1972).Onthe levelofspecies,thesizeoftheadenodactylincomparison withthepenispapilla,andviceversa,formsauseful taxonomiccharacter. Adenodactylsorcomparativestructureshavebeen reportedforaconsiderablenumberofplanarians,includingland,freshwater,andmarinespecies.Morphological differencessuggestthattheseorgansarenothomologous andevolvedindependentlyinvariousgroupsofspecies (SluysandRohde,1991,andreferencestherein).Adenodactylsoccurinseveraldendrocoelidgenera: Dendrocoelum Caspioplana Zabusova,1951, Baikalobia Kenk,1930, Polycladodes Steinmann,1910,and Acromyadenium De Beauchamp,1931.Ouranalysisisrestrictedtothegenus Dendrocoelum ,thuslikelyinvolvinghomologousadenodactyls. Within Dendrocoelum ,adenodactylsarecone-shaped structureswithadistinctbulbandpapilla,providedwitha well-developedlumenthatreceivestheabundantsecretion ofmanyglands.Thepreciselocation,orientation,and shapeoftheadenodactylvarybetweenspeciesandform importanttaxonomiccharactersthathavebeentakeninto considerationbytheclassicalstudiesontheseanimals. However,thesamestudiesconsistentlyfailtodocument theprecisearrangementofthemusculatureofthe adenodactyls(e.g.,Stankovic andKoma rek,1927;De Beauchamp,1932;Kenk,1978),thusapparentlyechoeing VonGraff’s(1912–17:3112)statementthatinfreshwater planariansalladenodactylsarehistologicallyalike. However,inthepresentstudyandotherjointprojects, itwasnoticedthatparticularlythe Dendrocoelum species fromtheLakeOhridregionhaveahistologicallycharacteristictypeofadenodactyl.Themoststrikingfeatureof theseadenodactylsisthepresenceofazoneoffinecircular musclefibersthatrunsthroughthemesenchymeof thepapillaandstainsbrightbluewhenstainedwitha trichromestainsuchasMallory-Cason.Thisbluezone runsthroughthepapillaandextendsontothebulbofthe adenodactyl.Ectallyandentallytothisbluezoneof circularmusclesrunlayersoflongitudinalfibers,which stainred,asdoallothermusclefibersintheanimal. Especiallyin Dendrocoelumadenodactylosum thissituation canbeobservedclearlybecauseoftheverylargesizeofthe adenodactyl(Fig.12).Weobservedthesameadenodactyl histologyin Dendrocoelummaculatum (Stankovic and Koma rek,1927)(Fig.13)andexpectthebluezoneoffine circularmusclestobepresentintheadenodactylofmany otherspeciesfromtheBalkanregion.Wewillcallthisthe Balkantypeofadenodactyl,albeitthatitsoccurrenceisnot restrictedtothisgeographicregion.Forexample,itisalso presentinthewidelydistributed Dendrocoelumlacteum (Mu ¨ller,1774)(Fig.14).TheBalkantypeofadenodactylis alsopresentinthethreenewTunisianspeciesdescribed aboveandinsomespeciesfromSardinia(Stocchinoetal., inprep.).Itistruethatintrichromestains,musclefibers alwaysstaineitherblueishorreddish,eveninthesame preparation.However,thedifferencewiththesespeciesof Dendrocoelum isthatthereissuchacleardifferencein stainingpropertiesofthebluezoneofcircularmusclesof theadenodactylandthemusculatureoftherestofthe body.Forexample,inaseriesofsectionsof Dendrocoelum amplum fromTunisiathatgenerallyshowpooraffinityfor theanilineblue(Wasserblau)componentoftheMalloryCasonstain,thefinecircularmusclezoneofthe adenodactyl,nevertheless,stainsbrightlightblue.The chemicalreasonwhythisparticularzoneofmusclesstains somuchdifferentlyremainsanenigma.Reexaminationof thematerialof D.spatiosum Vila-Farre andSluys,2011 fromnortheasternSpainrevealedthatthethickcircular zoneofmesenchymalmusclesisalsopresentinthe adenodactylofthisspecies,althoughitsdifferential stainingismuchlessexplicitthanintheotherspecies mentioned. TheBalkantypeofadenodactylisabsentinspeciessuch as Dendrocoelumnausicaae Schmidt,1861and D.beauchampi DelPapa,1952.Insuchspeciesthemesenchymal musculatureoftheadenodactylconsistspredominantlyof A.H.H ARRATH ,R.S LUYS ,A.G HLALA,AND S.A LWASEL JournalofCaveandKarstStudies, April2012 N 55

PAGE 58

longitudinalmuscles,withonlyathinlayerofcircular musclesdirectlyunderneaththeliningepitheliumofthe lumenoftheadenodactylthatstainsnodifferentlyfromthe othermusclefibers,asisthecasein D.nausicaae (Fig.15). In D.beauchampi, themesenchymalmusculatureofthe adenodactylpapillaispoorlydeveloped,merelyconsisting ofsomecircularfibers(SluysandBenazzi1992). Evidently,theprecisetaxonomicdistributionofthe Balkantypeofadenodactylneedstobestudiedand evaluatedinmoredetailinfuturestudies.However,inour opinionpresentevidencealreadysuggeststhatpresenceof thisstructuresignalsacloselyrelatedgroupof Dendrocoelum species,whichshowsageographicallyinteresting distributionthatcurrentlycomprisesnorthernEurope, southeasternEurope,Sardinia,northwesternAfricaand, mostlikely,alsonortheasternSpain. A CKNOWLEDGEMENTS TheauthorsextendtheirappreciationtotheDeanshipof ScientificResearchatKingSaudUniversityforfunding theworkthroughtheresearchgroupprojectNoRGP-VPP164. WearegratefultoMiquelVila-Farre andGiacinta StocchinoforhavingmadeavailabletoRSpreparationsof dendrocoelidsfromSpainandSardiniaandforsharing withustheirinformationontheanatomyoftheseanimals. MiquelVila-Farre isalsothankedformakingavailableand sectioningspecimensof Dendrocoelumlacteum fromthe EbrodeltainSpain.Prof.Dr.M.Kawakatsuisthanked fornomenclaturaladvice.A.H.Harrathisindebtedto Prof.FathiaZghalandProf.SaidaTekayafortheir supportandencouragement.Thisworkisdedicatedtothe Figures12–15.12. Dendrocoelumadenodactylosum fromnearResen(Crusje),Macedonia.Photomicrographofadenodactyl, showingtheconspicuous,blue-stainingzoneofcircularmuscle.13. Dendrocoelummaculatum fromBej-Bunarstreamnear Ohrid,Macedonia.Photomicrographofadenodactyl,showingtheconspicuous,blue-stainingzoneofcircularmuscle.14. Dendrocoelumlacteum fromtheEbrodelta,Spain.Photomicrographofadenodactyl,showingtheconspicuous,blue-staining zoneofcircularmuscle.15. Dendrocoelumnausicaae fromCorfu,Greece.Photomicrographofadenodactyl.Seecaptionto Figures6–11forkey. T HEFIRSTSUBTERRANEANFRESHWATERPLANARIANSFROM N ORTH A FRICA,WITHANANALYSISOFADENODACTYLSTRUCTUREINTHEGENUS D ENDROCOELUM (P LATYHELMINTHES ,T RICLADIDA ,D ENDROCOELIDAE ) 56 N JournalofCaveandKarstStudies, April2012

PAGE 59

memoryofFattouma,themotherofAHH,withoutwhom hewouldneverhavebecomethepersonthatheisatpresent. AHHisalsogratefulforthesupportofhisbelovedwifeM. JallouliHarrath,whoreinvigoratedhistasteoflife. R EFERENCES Ball,I.R.,andReynoldson,T.B.,1981,BritishPlanarians:Cambridge, CambridgeUniversityPress,141p. DeBeauchamp,P.,1932,BiospeologicaLVI,Turbellarie s,Hirudine es, Branchiobdellide s,Deuxie `meSe rie:ArchivesdeZoologieExpe rimentaleetGe ne rale,v.73,p.113–380. DeBeauchamp,P.,1937,Turbellarie stricladesdeYougoslaviere colte s parMM.RemyetHubault:BulletindelaSocie te Zoologiquede France,v.62,p.351–365. DeBeauchamp,P.,1954,Nouvellesdiagnosesdetricladesobscuricoles VIII-IX:BulletindelaSocie te ZoologiquedeFrance,v.79, p.418–427. DelPapa,R.,1965,Descrizionedi Dendrocoelum ( Eudendrocoelum ) botosaneanii n.sp.dellegrottedelBanato(Romania):Monitore ZoologicoItaliano,v.73,p.156–162. Ghlala,A.,DellaValle,D.,andMessana,G.,2009,Firstrecordofthe genus Typhlocirolana Racovitza,1905(Isopoda:Cirolanidae)from TunisiaanddescriptionofanewspeciesfromtheNationalParkof Ichkeul:Zootaxa,no.2176,p.57–64. Gourbault,N.,1972,RecherchessurlesTricladespaludicoleshypoge s: Me moiresduMuse umNationald’HistoireNaturelle,nouvellese rie (Se rieA),Zoologie73,249p. Jeannel,R.,andRacovitza,E.,1914,BiospeologicaXXXIII:Enume rationdesgrottesvisite es1911–1913:ArchivesdeZoologieExpe rimentaleetGe ne rale,v.53,p.325–558. Juberthie,C.,Decu,V.,Aellen,V.,andStrinati,P.,2001,Tunisie, in Juberthie,C.,andDecu,V.,eds.,EncyclopaediaBiospeologica,Tome III:Bucarest,Socie te InternationaledeBiospe ologie,p.1719–1728. Kenk,R.,1978,Theplanarians(Turbellaria:TricladidaPaludicola)of LakeOhridinMacedonia:SmithsonianContributionstoZoology, no.280,56p. Messana,G.,2004,AfricaBiospeleology, in Gunn,J.,ed.,Encyclopediaof CavesandKarstScience:NewYorkandLondon,FitzroyDearborn, p.24–25. Porfirjeva,N.A.,1977,PlanariiOzeraBaikal:Novosibirsk,NaukaPublishing,206p.[inRussian] Seurat,L.G.,1921,FaunedeseauxcontinentalesdelaBerbe rie:Alger, PublicationsdelaFaculte desSciences,TravauxduLaboratoirede Zoologieapplique e,66p. Seurat,L.G.,1934,Fauneaquatiquedusudetdel’extre ˆmesuddela Tunisie:AnnalesdesSciencesNaturelles,Zoologie,v.17,p.441–450. Sluys,R.,2007,Annotationsonfreshwaterplanarians(Platyhelminthes TricladidaDugesiidae)fromtheAfrotropicalRegion:Tropical Zoology,v.20,p.229–257. Sluys,R.,andBenazzi,M.,1992,Anewfindingofasubterranean dendrocoelidflatwormfromItaly(Platyhelminthes,Tricladida, Paludicola):Stygologia,v.7,p.213–217. Sluys,R.,andKawakatsu,M.,2006,Towardsaphylogeneticclassificationofdendrocoelidfreshwaterplanarians(Platyhelminthes):a morphologicalandeclecticapproach:JournalofZoologicalSystematicsandEvolutionaryResearch,v.44,p.274–284.doi:10.1111/ j.1439-0469.2006.00371.x. Sluys,R.,andRohde,K.,1991,Anewspeciesoffreshwatertriclad (Platyhelminthes:Tricladida)fromAustralia:ZoologicalJournalof theLinneanSociety,v.102,p.153–162.doi:10.1111/j.10963642.1991.tb00286.x. Stankovic ,S.,andKoma rek,J.,1927,DieSu ¨wasser-Tricladendes WestbalkansunddiezoogeographischenProblemedieserGegend: ZoologischeJahrbu ¨cherAbteilungfu ¨rSystematik,O ¨ kologieund GeographiederTiere,v.53,p.591–674. Stocchino,G.A.,Sluys,R.,Manconi,R.,Casale,A.,Marcia,P.,Grafitti, G.,Cadeddu,B.,Corso,C.,andPala,M.,(inprep.),Tricladsfrom Sardiniangroundwaters. Vila-Farre ,M.,Sluys,R.,Almagro,I .,Handberg-Thorsager,M.,and Romero,R.,2011,Freshwaterplanarians(Platyhelminthes,Tricladida)fromtheIberianPeninsulaandGreece:diversityandnoteson ecology:Zootaxa,no.2779,p.1–38. VonGraff,L.,1912–17,Tricladida, in Dr.H.G.Bronn’sKlassenund OrdnungendesTier-Reichs,Bd.IVVermes,Abt.IC:Turbellaria,II Abt.:Tricladida:Leipzig,C.F.Winter,p.2601–3369. A.H.H ARRATH ,R.S LUYS ,A.G HLALA,AND S.A LWASEL JournalofCaveandKarstStudies, April2012 N 57

PAGE 60

MICRO-CHARCOALABUNDANCESINSTREAM SEDIMENTSFROMBUCKEYECREEKCAVE,WEST VIRGINIA,USA G REGORY S.S PRINGER 1 ,L.N IVANTHI M IHINDUKULASOORIYA 2 ,D.M ATTHEW W HITE 3 AND H AROLD D.R OWE 4 Abstract: Wecomparemicro-charcoalabundancesinlaminatedcave-streamsediments tothepresencesofNativeAmericansandlatersettlersinthesamewatershed.Samples wereobtainedfromacoretakenfroma2.5mhighpointbarlocated1kminsideof BuckeyeCreekCave,WestVirginia.Thirty-threesubsamplesweretreatedwithhydrogen peroxidetobleachorwhitennon-charcoalorganicmatter.Intheabsenceofopaque mineralgrains,thistechniquecreatesalargevisualcontrastbetweendarkcharcoalgrains andothersubstances.Thesubsampleswerephotographedusingamicroscope-mounted camera,andpixelsdarkerthan99/255(grayscale)wereusedtocalculatecharcoal concentrations.Therecordspansthelast6,000years,andfourofthefivehighestcharcoal concentrationsarefromthelast2,000years.ThehighestconcentrationisfromAD1093, andthesecond-highestconcentrationisfromthenineteenthcentury.Post-Colonialsettlers beganmakingextensiveuseofthewatershedsometimeintheeighteenthcenturyandmay, therefore,beresponsibleforthesecond-highestcharcoalconcentration.However, archaeologistsindependentlyconcludedthatNativeAmericansmadepeakuseofthe watershedbetweenAD1000and1200,whichcoincideswiththehighestcharcoal concentrationintherecord.NativeAmericansareknowntohaveextensivelyusedfire,so thereisgoodcircumstantialevidencetyinghighconcentrationsinthelast2,000yearsto humanactivities.Ourmethodissuitableforuseelsewhere,andwepresentadetailed statisticalanalysisofourdataasaguidetowardinterpretingcharcoalconcentrationsin karstandnon-karstdeposits. I NTRODUCTION Clasticcavesedimentspreserverecordsofgeomorphic processes,climate,andlandusesasfluctuationsin geochemistry,mineralogy,ororganicmatter(Springer, 2005).Thelatterincludescharcoal,theabundanceofwhich isanespeciallyvaluableproxyforlanduse.However, sedimentary-charcoalabundancesarelittlestudiedintrue caves,asopposedtorockshelters.Charcoalabundances arewidelymeasuredinlacustrinesediments,andlakebasedstudiesoftenfocusonreconstructingdetailedfire historiesandidentifyingpeaktimesofcharcoaldeposition (Higueraetal.,2007).Thesehavemanyapplications withinclimate,forestsuccession,andarchaeological studies(Conederaetal.,2009;The ry-Parisotetal.,2010). Conceptually,highcharcoalabundancesmayrepresent periodsoffrequentfiresorgreaterfireseverities,but mixingandstorageduringtransportthroughstream networkscreatesapersistentbackgroundsignal,while themechanismsoftransportandthedistancesoverwhich charcoalistransportedsignificantlyaffectparticlesizeand abundanceinsedimentarydeposits(Conederaetal.,2009; Higueraetal.,2007).Theselimitationsexplainwhymuch attentionisgiventopeaksincharcoalabundancedata. Nonetheless,thefluvialprocessesthateffectivelyfilterand smoothcharcoalabundancesarepresumablytrueofboth surfaceandsubsurfacestreamsandcharcoalabundancesin clasticcavesedimentsmaybevaluableforunderstandinga varietyofnaturalandanthropogenicprocesses(Carcaillet etal.,2007). Aspartofageoarchaeologicalstudy,wesoughtto efficientlymeasureastatisticallymeaningfulnumberof charcoalabundancesinclasticcavesediments(n $ 30)by refininganexistingcharcoalseparationmethod(Rhodes, 1998).Weexaminedfine-grainedsedimentscontaining onlymicro-charcoalfragmentslessthan1mmindiameter andnotreadilyvisibletotheunaidedeye.Afterseparating ordistinguishingcharcoalfromunconsolidatedsediments andotherdetritus,thenecessarydataareobtainedby countingcharcoalgrainsormeasuringcharcoal-grain surfaceareasorvolumes(Alietal.,2009).Theseparation andquantificationprocedurescanbeverytimeconsuming andimpactsthenumberandspacingofsamplesandtheir temporalresolution(Rhodes,1998).Theseissuescanlimit *CorrespondingAuthor 1 DepartmentofGeologicalSciences,OhioUniversity,Athens,OH45701springeg@ ohio.edu 2 DepartmentofGeology,KentStateUniversity,Kent,OH45701lnivanthi@ yahoo.com 3 ExxonMobil,800BellStreet,Room2897D,Houston,TX77002david.m.white@ exxonmobil.com 4 DepartmentofEarthandEnvironmentalSciences,TheUniversityofTexasat Arlington,Arlington,TX76019hrowe@uta.edu G.S.Springer,L.N.Mihindukulasooriya,D.M.White,andH.D.Rowe–Micro-charcoalabundancesinstreamsedimentsfromBuckeye CreekCave,WestVirginia,USA. JournalofCaveandKarstStudies, v.74,no.1,p.58–64.DOI:10.4311/2010AN0148R1 58 N JournalofCaveandKarstStudies, April2012

PAGE 61

aresearcherÂ’sabilitytodistinguishbetweencharcoalabundancevaluesofanthropogenicandnon-anthropogenicorigins.Therefore,weuseddigitalprocessingtogenerate datamoreefficientlyandinsufficientnumbersfor statisticalanalysis.Wedescribeourlaboratorymethod andprovidestatisticalanalysisofourresults.The reliabilityofthelatteristestedbybrieflycomparing micro-charcoalabundancestoindependentgeological (Springeretal.,2008)andarchaeologicaldata(McBride andSherwood,2006;Springeretal.,2010).Indoingso,we assumedthatcharcoalabundancesreflectsomeaspectof firefrequenciesorseverities(Higueraetal.,2007;Conedera etal.,2009;The ry-Parisotetal.,2010),butdonotinterpret ourrecordasadetailedfirehistory.Instead,wefocuson peakabundancesandwhetherthesepeaksarecorrelated withhumanactivitiesinthesourcewatershed.Ourmicrocharcoaltechniqueshouldbeofvaluetoothersandhas manypotentialapplicationsincaveandkarststudies. S TUDY A REA Wereportdatacollectedfromapointbardepositedbya streaminBuckeyeCreekCave,whichdrainsatopographicallyenclosed,14km 2 watershedinsoutheasternWest Virginia,USA(Fig.1).BuckeyeCreekflowsthrougha2km longpassageaveraging10mwideand4mhigh(Dasherand Balfour,1994;Springer,2004).Siltbanksarefoundonone orbothsidesofthestreamformuchofitsunderground course.Thefine-grainedsedimentsarederivedfromakarst watersheddrainingMississippian-agesandstones,siltstones, andshalesofButlerMountain.Surfacestreamscarry siliciclasticdetritustothebaseofButlerMountain,where streamssinkintotheirbedsorswalletholesinlimestoneflooredvalleys(Springeretal.,2003). M ETHODS Wedugatrenchinthefaceofa2.6mhighpointbar, 1kmdownstreamofthecaveentranceina20mwideby 7mhighpassage.Depositioncontinuesatopthepointbar, althoughthestreamhasmeanderedagainstthebarto createacutbank.Thesedimentslackinteriorscour surfaces,hard-grounds,andotherindicationsoferosion orperiodsofnon-deposition.Therefore,weassume depositionhasbeenessentiallycontinuoussincethe sedimentsbeganaccumulating.Thesedimentsweredescribedandsampledinthefield,butwealsocoredtheface ofthetrenchbypushingPVCsleevesintothesediments. ThesleevesaremadefromPVCtubeswithsquarecross sections,measuring10.2cmonaside(4inchessquare),cut lengthwisetoyieldasquare-corneredtrough.Corelengths weretypically 50cm,andthesleevesoverlappedtocreate acontinuoussampleofthesediments.Theopensidesof sleeveswerepressedintosedimentbeforebeingcutoutof thetrenchwall,wrappedinprotectivematerials,and carriedoutofthecave.Thisallowedustoperformdetailed subsamplingandlaboratoryanalysesinacontrolled environment. Thecoresweresampledatselecteddepthsforcharcoal analysisusingamethoddescribedbyRhodes(1998). Subsamplesweighing2gwereremovedfromsediment coresanddriedfortwodays,rehydratedwithdilute potassiumhydroxide(KOH),decanted,andtreatedwith 20mlof5%hydrogenperoxide(H 2 O 2 ).Theperoxide treatmentbleachesallorganicmatterexceptcharcoal, whichremainsdarkandcontrastsmarkedlywiththe bleachedorganicsandlight-coloredandtranslucent sedimentgrains.Weusedthecontrastcreatedbythe H 2 O 2 treatmenttoestimatecharcoalconcentrationsusing adigitaltechnique.Thepreparedsampleswereplacedin petridishesandphotographedusingdigitalphotomicroscopy.Individualimageswereconvertedtograyscale,and allpixelsbelowagrayscalethresholdof99/255were countedusingImageJsoftware(http://rsbweb.nih.gov/ij/) (Fig.2).Theuseofpixel-countinggreatlyspeedsestimationofcharcoalconcentrations,butitover-countscharcoal abundancesbecausesomedarkobjectsareprobablynot charcoal.However,wevisuallyscannedalldishesduring Figure1.BuckeyeCreekislocatedinsoutheasternWestVirginia(leftpanel)andentersthecaveafterflowingacrossan activefloodplainwithinalargekarstdepression(rightpanel).Aplanviewofthecaveissuperimposedonbasintopographyat right,andthelocationofoursedimenttrenchisindicated.AdaptedfromDasherandBalfour(1994)andSpringeretal.(2010). G.S.S PRINGER ,L.N.M IHINDUKULASOORIYA ,D.M.W HITE AND H.D.R OWE JournalofCaveandKarstStudies, April2012 N 59

PAGE 62

photomicroscopyanddidnotobserveanydarkmineral grains,probablybecausethesiltsandsandsaregreater than99%quartz,asmeasuredusingX-raydiffraction. However,tocontrolforcontaminationfromdark,noncharcoalparticles,weaveragedtheresultsofmultiple imagestakenfromdifferentareaswithinindividualpetri dishes. M ICRO -C HARCOAL R ESULTSAND S TATISTICS Atotalof33sampleswereanalyzedformicro-charcoal concentrations.Thefirstsamplewastaken1cmbelowthe topofthedeposit(core),andadditionalsampleswere takenevery2cmtoadepthof27cm(total n 5 14). Thereafter,sampleswerecollectedevery3to4cmtoa depthof98cm( n 5 19).Ageswereassignedtoeachsample usinganage-depthmodelbuiltfromsevenradiocarbon dates.Themodelanditsuncertaintiesarediscussedin Springeretal.(2010),andthebasalsedimentsare 6,500yearsold.Charcoalconcentrationsrangeoveran orderofmagnitude(0.22to2.15mm 2 cm 2 2 ).Overall, valuesarehighestinthelateHolocene,butthereis considerablevariabilityduringthistime(Fig.3).Overall, concentrationsarestronglyskewedtotheleftandfaileda Shapiro-Wilktestfornormality( p 5 0.007).Usingthe sametestandabase-tenlogtransformation,thevaluesare lognormallydistributed( p 5 0.995)andtheircumulative frequenciesfallonanearlystraightlinewhenplottedusing aprobabilityscale(Fig.4).Nonetheless,thelowestand fivehighestpointsdonotfallontheline.Followingthe methodsofSinclair(1974)andReimannetal.(2005),we didnotperformoutliertests,butcalculatedthresholds abovewhichdatapointsareanomalousanddeserving furtheranalysis. Thresholdvaluescanbechoseninavarietyofways. Conventionalmethodsincludeusingthevaluecorrespondingtothemean 6 2standarddeviations(mean 6 2sdev), median 6 2medianabsolutedeviation(MAD),andthe box-plotmethod(Reimannetal.,2005).Accordingto mean 6 2sdevmethod,valuesoutsideupperandlower thresholdsareconsideredpossibleoutliersworthyof furtherattention.However,thismethodidentifiesonly 5%ofactualextremevalues(Reimannetal.,2005),andin ourcase,isexceededonlybytheAD1093value(Fig.4). Thebox-plotmethodwasperformedusingSystatÂ’s SigmaPlotandidentifiedtheAD1889,AD1093,and AD24valuesasoutliers(Fig.4inset).TheMADmethod usuallyresultedinalowerthresholdlevelthanthebox-plot ormean 6 2sdevmethods,anditsthresholdwaslower thanthefourlargestcharcoalconcentrations(Fig.4). Figure2.Charcoalconcentrationsweremeasuredbyphotographingpreparedsedimentsamples(A),convertingtheimagesto grayscale(B),andcountingallpixelsbelowagrayscalethresholdof99/255(C).Thedarkestpixelswereassumedtobe charcoalbecauseextensivepretreatmentwithH 2 O 2 bleachedallotherorganicmatterandnodarkmineralgrainswere observedduringphotographing. M ICRO-CHARCOALABUNDANCESINSTREAMSEDIMENTSFROM B UCKEYE C REEK C AVE ,W EST V IRGINIA ,USA 60 N JournalofCaveandKarstStudies, April2012

PAGE 63

However,thesemethodsdonottakeintoaccountthe possibilitythattheanomalousvaluesarederivedfroma differentpopulationthanthebackgroundpopulationof pointsfallingonanapproximatelystraightlinein Figure4.Iftwopopulationshavebeensampled,themean andstandarddeviationofthepooleddatamaynot bestatisticallymeaningful.Asaconsequence,thestrict applicationofthresholdvaluesyieldsquestionableresults (cf.Reimannetal.,2005). Therearealternativestousingthresholdsofpurely quantitativederivation.Amongthese,themostpopular remainssubjectiveinterpretationofnon-linearityincumulativeprobabilityplots,wherebythepointofmaximum curvatureisinterpretedasthethresholdseparating differentpopulations(Sinclair,1974;Reimannetal., 2005).Pointsbeyondthecurvaturethresholdshouldbe subjectedtoadditionalinvestigation,becauseMonte Carlosimulationsindicatethismethodyieldstoomany outliers(Reimannetal.,2005).Whenplottedascumulative probabilities(Fig.4),thefivehighestconcentrationsdo notfallonatrendlinedrawnthroughintermediate charcoalconcentrationvaluesandfivesamplesqualifyas beinganomalous:AD1889,AD1093,AD452,AD24, and1730BC. D ISCUSSION Thelast2,000yearsincludesfourofthefivehighest micro-charcoalconcentrationsandallstatisticalmethods identifiedtheAD1093eventasananomalousvalue.The charcoalisderivedfromaverysmallwatershed(14km 2 ), whichisidealbecauselargewatershedsallowtoomuch particlemixingandsmoothingofpeakfireevents (Carcailletetal.,2007).Variabilityinlocallygenerated charcoaldepositsusuallyreflectsanevent-dominated record,withmajorfireepisodesrecordedasspikesin charcoalabundances(Clarketal.,1996;Marlonetal., 2006;Carcailletetal.,2007).Alternatively,variabilitycan reflectoneormoreflawsinsamplingormeasurement methodologies.Onthismatter,wenotethatWhite(2007) containsalowerresolutionmicro-charcoalrecordmade usingthesametechniques.Hisrecordalsoshowshighlate Figure3.(A)Astableisotopic( 13 C calc )paleoclimaticrecordobtainedfromastalagmiterecordslong-termchangesinthesoils andecosystemaboveBuckeyeCreekCave(Springeretal.,2008).MoisturelevelsincreasedfromthemidtolateHolocene,but 13 C calc valuesincreasedabruptly 100BCThiscoincideswithuseofUnnamedCave # 14(UC14)byNativeAmericans,and valuespeakapproximatelywhenNativeAmericansweremakingtheirmostextensiveuseofUC14(Maslowski,2006).Other regionalpaleoclimaticrecordsdonotrecorddryingafter2,100yearsBP(Springeretal.,2009;2010),buttheanomalous 13 C calc recordiscorroboratedbystableisotopicanalysisoforganicmaterial( 13 C org )intheTrenchFivesediments(B).(C)Fire activitywashighduringtheLateHoloceneandpeaksnearthetimesofpeak 13 C calc and 13 C org values,coincidingwithpeak NativeAmericanuseofthewatershed,asinferredfromtheUC14record(McBrideandSherwood,2006). G.S.S PRINGER ,L.N.M IHINDUKULASOORIYA ,D.M.W HITE AND H.D.R OWE JournalofCaveandKarstStudies, April2012 N 61

PAGE 64

Holocenecharcoalabundancesrelativetomid-Holocene values.Nonetheless,thehighestconcentrationsareabove severalthresholds(Fig.4)and,therefore,worthadditional consideration.So,wecomparetheirtimingstoan independentarchaeologicalinvestigationconductedin UnnamedCave # 14(UC14)adjacenttoBuckeyeCreek Cave(Fig.5).(Thecommonnameandentrancelocationof UC14areonfilewiththeStateofWestVirginiaDivision ofCultureandHistory,butarewithheldtoprotect enclosedculturalremains.)Thecomparisonisatestof whetherhumanscouldberesponsibleforthehighcharcoal concentrations. UC14consistsofa30mby46mentranceroomand morethan100mofinteriorpassages.Theentranceroomis morethan3mhighandhospitableforhumanhabitation, withaflatflooranddaytimeilluminationfroma southeast-facingentrance.Drs.KimMcBrideandSarah SherwoodoftheKentuckyArchaeologicalSurveyand DickinsonCollegeinPennsylvania,respectively,ledthe projectandco-editedavolumecomposedofchapters dedicatedtostoneimplementsanddebitage,pottery sherds,petroglyphs,hearths,andfoodstuffs(McBride andSherwood,2006).ThereportisonfilewiththeWest VirginiaCaveConservancy(http://www.wvcc.net/). Culturalremainswereassignedagesbycomparingtheir morphologiestopublished,well-datedculturalmaterials fromotherregionalexcavations(Fig.5)(Maslowski,2006; McBride,2006;SimekandCressler,2006).Collectively,the culturalremainsrevealthatUC14wasutilizedinoneor morewaysfromtheLateArchaicthroughtheLate Prehistoric(5,000to250calendaryearsBP),atleast occasionally.Thethreemostculturallydistinctiveartifact classesareprojectilepoints,pottery,andpetroglyphs.All threetypesofremainsarerepresentedbetweenAD1000 and1200(750and550yearsBP;blackinFig.5). Maslowski(2006)inferredthatpeakNativeAmerican utilizationofUC14occurredduringthattimeperiod.This iscircumstantialevidencethatNativeAmericansare responsibleforfiresthatproducedseveralofthehighest charcoalconcentrations. Figure4.Charcoalconcentrationsdisplayalog-normaldistribution(inset).Thecumulativedistributionfunctionofcharcoal concentrationsisplotted,andfiveanomalouslyhighconcentrationsarelabeled.Thresholdscalculatedtoidentifyanomalous valuesareshownasdashedlinesandwithintheboxplot(inset)asblackcircles.Detailsdiscussedinthetext. M ICRO-CHARCOALABUNDANCESINSTREAMSEDIMENTSFROM B UCKEYE C REEK C AVE ,W EST V IRGINIA ,USA 62 N JournalofCaveandKarstStudies, April2012

PAGE 65

However,thesecondhighestvalueisfoundat1cm depthinTrenchFive.Thecorrespondingdate,AD1889,is afterEuropeansbeganmakingsubstantialchangesto GreenbrierCountyecosystems(Hale,1886).Knowingthat settlersandtheirdescendantsperformedextensiveland clearance(Hale,1886),itseemshighlyprobablethatthe second-highestpeakreflectshumanactivities.Afifthpoint, 1730BC,alsoliesabovethetrendline,butonlyafew ArchaicartifactswerefoundinUC14,andthesignificance ofthehighmicro-charcoalabundancevalueisunclear. C ONCLUSIONS Fourofthefivehighestmicro-charcoalabundancesin BuckeyeCreekCavesedimentsweregeneratedattimes whenhumansweremakingsignificantuseoftheBuckeye Creekwatershed.NativeAmericansareknowntohave managedforestsusingfire,andsubsequentsettlersused firetoclearland,sothereisstrongcircumstantialevidence thathumansareresponsibleforhighcharcoalconcentrationscreatedduringthelateHolocene(Springeretal., 2010).Archaeologicalstudiesoftenmakeuseofclastic sedimentsinrockshelters(Springer,2005),butcave interiordepositsarerarelyusedaswehavedone.Microcharcoalmayrepresentavaluableresourcebecauseit accumulatesinlaminatedsedimentswithsimpleorigins, andverysmallsamplesaresufficientforanalysis.Noncaverecordsmaybeunavailableinmanysettings,so agreementbetweenourmicro-charcoalresultsandhuman activitiesgiveconfidencethatmicro-charcoalincave sedimentscanbeusefulforreconstructinglandusesand ecologicalregimes.Werecommendmicro-charcoaldatabe subjectedtostatisticalanalysis,becausehighconcentrationsarenotuniquelyassociatedwithanyparticularsource (Conederaetal.,2009).Wealsobelievethreshold-based approachesusedingeochemicalstudiesareappropriatefor micro-charcoalstudies(Sinclair,1974;Reimannetal., 2005),butrecommendgeneratingasmanydatapointsasis practical. R EFERENCES Ali,A.A.,Higuera,P.E.,Bergeron,Y.,andCarcaillet,C.,2009, Comparingfire-historyinterpretationsbasedonarea,numberand estimatedvolumeofmacroscopiccharcoalinlakesediments: QuaternaryResearch,v.72,no.3,p.462–468.doi:10.1016/j.yqres. 2009.07.002. Carcaillet,C.,Perroux,A.,Genries,A.,andPerrette,Y.,2007, Sedimentarycharcoalpatterninakarsticundergroundlake,Vercors massif,FrenchAlps:implicationsforpalaeo-firehistory:The Holocene,v.17,no.6,p.845–850.doi:10.1177/0959683607080524. Clark,J.,Royall,P.,andChumbley,C.,1996,Theroleoffireduring climatechangeinaneasterndeciduousforestatDevil’sBathtub,New York:Ecology,v.77,no.7,p.2148–2166.doi:10.2307/2265709. Conedera,M.,Tinner,W.,Neff,C.,Meurer,M.,Dickens,A.F.,and Krebs,P.,2009,Reconstructingpastfireregimes:methods,applications,andrelevancetofiremanagementandconservation:Quaternary ScienceReviews,v.28,no.5–6,p.555–576.doi:10.1016/j.quascirev. 2008.11.005. Dasher,G.,andBalfour,W.,eds.,1994,TheCavesandKarstofthe BuckeyeCreekBasin:Barracksville,WestVirginia,WestVirginia SpeleologicalSurveybulletin12,326p. Hale,J.,1886,Trans-AlleghenyPioneers:HistoricalSketchesoftheFirst WhiteSettlementsWestoftheAlleghenies1748andAfter:Cincinnati, Ohio,GraphicPress,330p. Higuera,P.E.,Peters,M.E.,Brubaker,L.B.,andGavin,D.G.,2007, Understandingtheoriginandanalysisofsediment-charcoalrecords withasimulationmodel:QuaternaryScienceReviews,v.26,no.13– 14,p.1790–1809.doi:10.1016/j.quascirev.2007.03.010. Marlon,J.,Bartlein,P.J.,andWhitlock,C.,2006,Fire-fuel-climate linkagesinthenorthwesternUSAduringtheHolocene:The Holocene,v.16,no.8,p.1059–1071.doi:10.1177/0959683606069396. Maslowski,R.,2006,Ceramicanalysis, in McBride,K.,andSherwood,S., eds.,ReportofArchaeologicalInvestigationsat[UnnamedCave # 14],WestVirginia:Lexington,Kentucky,KentuckyArchaeological Survey,p.7.1–7.10. McBride,J.,2006,Analysisofstonetools, in McBride,K.,andSherwood, S.,eds.,ReportofArchaeologicalInvestigationsat[UnnamedCave # 14],WestVirginia:Lexington,Kentucky,KentuckyArchaeological Survey,p.8.1–8.9. McBride,K.,andSherwood,S.,eds.,2006,ReportofArchaeological Investigationsat[UnnamedCave # 14],WestVirginia:Lexington, Kentucky,KentuckyArchaeologicalSurvey. Reimann,C.,Filzmoser,P.,andGarrett,R.G.,2005,Backgroundand threshold:criticalcomparisonofmethodsofdetermination:Scienceof TheTotalEnvironment,v.346,no.1–3,p.1–16.doi:10.1016/j.scitotenv. 2004.11.023. Rhodes,A.N.,1998,Amethodforthep reparationandquantificationof microscopiccharcoalfromterrestrialandlacustrinesedimentcores:The Holocene,v.8,no.1,p.113–117.doi:110.1177/095968369800800114. Figure5.Thegrayregionsrepresentculturalagesassigned toprojectilepoints(McBride,2006),pottery(Maslowski, 2006),andpetroglyphs(SimekandCressler,2006)foundin UnnamedCave # 14.Theblackbarinthepotterycolumn denotestheagerangeofdistinctivePageCordceramics (Maslowski,2006).Theblackstarisaradiocarbondate obtainedfrompotteryfragments(McBrideandSherwood, 2006). G.S.S PRINGER ,L.N.M IHINDUKULASOORIYA ,D.M.W HITE AND H.D.R OWE JournalofCaveandKarstStudies, April2012 N 63

PAGE 66

Simek,J.,andCressler,A.,2006,Prehistoriccaveartin[UnnamedCave # 14],WestVirginia, in McBride,K.,andSherwood,S.,eds.,Report ofArchaeologicalInvestigationsat[UnnamedCave # 14],West Virginia:Lexington,Kentucky,KentuckyArchaeologicalSurvey, p.3.1–3.13. Sinclair,A.,1974,Selectionofthresholdvaluesingeochemicaldatausing probabilitygraphs:JournalofGeochemicalExploration,v.3,no.2, p.129–149.doi:10.1016/0375-6742(74)90030-2. Springer,G.S.,2004,Apipe-based,firstapproachtomodelingclosed conduitflowincaves:JournalofHydrology,v.289,no.1–4, p.178–189.doi:10.1016/j.jhydrol.2003.11.020. Springer,G.S.,2005,Clasticsedimentsincaves, in Culver,D.,andWhite, W.,eds.,EncyclopediaofCaves:Amsterdam,Boston,Academic Press,p.102–108. Springer,G.,Rowe,H.,Hardt,B.,Edwards,R.L.,andCheng,H.,2008, SolarforcingofHolocenedroughtsinastalagmiterecordfromWest Virginiaineast-centralNorthAmerica:GeophysicalResearchLetters, v.35,L17703p.doi:10.1029/2008GL034971. Springer,G.S.,Rowe,H.,Hardt,B.,Cocina,F.G.,Edwards,R.L.,and Cheng,H.,2009,Climatedrivenchangesinriverchannelmorphology andbaselevelduringtheHoloceneandLatePleistoceneof SoutheasternWestVirginia:JournalofCaveandKarstStudies, v.71,no.2,p.121–129. Springer,G.S.,White,D.M.,Rowe,H.D.,Hardt,B.,Mihindukulasooriya,L.N.,Cheng,H.,andEdwards,R.L.,2010,Multiproxyevidence fromcavesofNativeAmericansalteringtheoverlyinglandscape duringthelateHoloceneofeast-centralNorthAmerica:The Holocene,v.20,no.2,p.275–283.doi:10.1177/0959683609350395. Springer,G.S.,Wohl,E.E.,Foster,J.A.,andBoyer,D.G.,2003,Testing forreach-scaleadjustmentsofhydraulicvariablestosolubleand insolublestrata:BuckeyeCreekandGreenbrierRiver,WestVirginia: Geomorphology,v.56,no.1–2,p.201–217.doi:10.1016/S0169555X(03)00079-5. The ry-Parisot,I.,Chabal,L.,andChrzavzez,J.,2010,Anthracologyand taphonomy,fromwoodgatheringtocharcoalanalysis.Areviewof thetaphonomicprocessesmodifyingcharcoalassemblages,in archaeologicalcontexts:Palaeogeography,Palaeoclimatology,Palaeoecology,v.291,no.1–2,p.142–153.doi:10.1016/j.palaeo.2009. 09.016. White,D.M.,2007,ReconstructionandAnalysisofNativeAmericanland useduringthelateHolocene[M.S.Thesis]:Athens,Ohio,Ohio University.164p. M ICRO-CHARCOALABUNDANCESINSTREAMSEDIMENTSFROM B UCKEYE C REEK C AVE ,W EST V IRGINIA ,USA 64 N JournalofCaveandKarstStudies, April2012

PAGE 67

RESPONSEOFTHEKARSTPHREATICZONETOFLOOD EVENTSINAMAJORRIVER(BOHEMIANKARST,CZECH REPUBLIC)ANDITSIMPLICATIONFORCAVEGENESIS H ELENA V YSOKA 1 ,J IR I B RUTHANS 1,2 ,K AREL Z A K 3 AND J IR I M LS 1 Abstract: Hydraulicandhydrochemicalrelationshipsbetweenamediumgradientriver andakarstaquiferwerestudiedbywaterlevelandtemperatureloggingcombinedwith watergeochemistryand d 13 C.Thecavelakesareseparatedfromtheriverbyafloodplain upto150mwideformedbyagravelandsandlayerupto13mthickcoveredwithfinegrainedfloodplainsediments.Duringminordischargepeaks(waterlevelintheriver 1.5mabovethenormalriverstage),awaterleveloscillationinthecavelakessituated 40to190mawayfromriverisinducedbytheriverleveloscillation,buttheriverwater doesnotenteranyofthelakes.Thegroundwaterchemistryinthecavelakesdiffersfrom thatoftheriverwater.Lowbicarbonatecontentandhigh d 13 Cvaluesindicatethatsome ofcavelakes’watershaveundergoneCO 2 degassingandcalciteprecipitation.Duringa majorflood(recurrenceinterval 100years,levelrising7mabovethenormalstage), theriverwaterrapidlyfloodedthecavesthroughopeningsintherivercanyon(floodflowinjection),whilethoseconnectedtotheriverviaalluviumonlywerefloodedbyan elevatedgroundwaterstage,andtheresultingwaterlevelrisewasonlyabout50percent oftheriverlevelincrease.Asimplehydraulicmodelwassuccessfullyusedtosimulate andexplainthewatertableoscillationsinthecavelakes.Flood-flowinjectionhas recentlybeensubstantiallyreducedbylow-permeability,fine-grainedlateHolocene fluvialsedimentsthatcapcoarsegravelsintheriverfloodplain.Fastspeleogenesisby floodinjectionwouldbeexpectedinperiodswhentherivercanyonwasbareorfilledby gravelalone(glacialperiods,transitiontoHolocene).Icejamscausinglocalincreasesin theriverlevelarerecognizedasoneoffactorsthatcanbeimportantinspeleogenesis. I NTRODUCTION Generally,directinvasionofriverfloodwatersintothe karstenvironmentenablesfastdevelopmentofcaves(Palmer, 1991),sincewaterinriversoriginatingoutsideofthekarstarea canbeundersaturatedwithrespecttocalciumcarbonate,and themixingofriverwaterandtheusualkarstCa-HCO { 3 watersisanotherfactorenhancingcorrosionofthelimestone (Klimchouketal.,2000).Duringthe lastfewdecades,injection ofriverfloodwatersintokarstporositybegantobeconsidered asanimportantspeleogeneticprocess(Palmer,1991). ChoquetteandPray(1970)subdividedtheevolutionof carbonaterocksintothreetimeorporositystagesreflecting therockcycle.Depositionandearlyexposureare eogenetic;deepburialismesogenetic;postburialexposure anderosionaretelogenetic.Becausekarstificationresults fromprocessesnearthesurface,karstcanbesubdivided intotwomaintypes,eogeneticandtelogenetic(Vacherand Mylroie,2002).Telogeneticaquifersdifferfromeogenetic aquifersnotonlyintheirmuchlowerprimaryporosityand matrixpermeability(FloreaandVacher,2006),butalsoin thesteeperinclinationoftheirwatertablelevel(e.g.,BaillyComteetal.,2010).Hydraulicandhydrochemicalrelationshipsbetweenriversandeogenetickarstaquiferswere studiedbyKatzetal.(1998),MartinandDean(2001), Opsahletal.(2007)andothers. Intelogenetickarstaquifers,theoverwhelmingmajority ofstudiesconcerningtherelationshipbetweenallogenic streamsandkarstaquifersarefocusedonpartlyorfully sinkingstreams(e.g.Bailly-Comteetal.,2009;Doctor etal.,2006).Thereisalackofstudiesconcerningtheeffect ofariverthatcrossesthekarstareabutdoesnotlosewater intothekarstaquiferundernormalwaterstages. Thehydraulicandhydrochemicalrelationshipsbetween majorlowgradientormediumgradientrivers( # 1m/km inthispaper)crossingtelogenetickarstareasandthe phreaticzoneinkarstifiedlimestonearoundtheriver representatopicthatisnotyetfullyunderstood.The valleybottomsofmediumandlowgradientriversare typicallyfilledwithfluvialriversediments(Andersonand Anderson,2010)andthesedimentthicknessisaffectedby climaticchanges(e.g.,Springeretal.,2009).Hydraulic conductivityandwaterexchangebetweentheriverandthe karsticphreaticzonemaybeaffectedbythesedimentary valleyfill. *CorrespondingAuthor 1 FacultyofScience,CharlesUniversityinPrague,Albertov6,12843Praha2,Czech Republic.helenavysoka@hotmail.com 2 CzechGeologicalSurvey,Kla rov3,11821Praha1,CzechRepublic 3 InstituteofGeology,AcademyofSciencesoftheCzechRepublic,v.v.i.,Rozvojova 269,16502Praha6-Lysolaje,CzechRepublic H.Vysoka ,J.Bruthans,K.Z a k,J.Mls–Responseofthekarstphreaticzonetofloodeventsinamajorriver(BohemianKarst,Czech Republic)anditsimplicationforcavegenesis. JournalofCaveandKarstStudies, v.74,no.1,p.65–81.DOI:10.4311/2010ES0178R JournalofCaveandKarstStudies, April2012 N 65

PAGE 68

Whereriversfreezeinthewinter,theformationof eitherfrazil-icejamsduringriverfreezingormajoricejams duringthespringsnowmeltleadstotheformationofriver stretcheswithverysteepgradients.Differencesinthewater levelaboveandbelowtheicebarriercanbeashighas severalmeters(Matous ek,2004).Theseeventsgeneratea steephydraulicgradientbetweentheriverandthe limestoneaquifer. TheBohemianKarst,locatedinthecentralpartofthe CzechRepublicsouthwestofPrague,isasmallkarstarea thatishighlysuitableforstudyingtheserelationships.The karstregioniscrosscutbyacanyonoftheBerounkaRiver inasectionabout8kmlong.TheBerounkaRiverisa major,medium-gradient(0.79mperkm)riverdrainingthe westernpartoftheCzechRepublic.Altogethersixteen caveshavebeenfoundaroundtheriverthateithercontain smallpermanentcavelakesorwherewateroccursperiodicallyduringriverfloods.Thisrivervalleysuffereda majorfloodinAugust2002(afloodwithrecurrence intervalof500to1000years;datafromwww.chmi.cz, MoECR,2003).Observationsofthebehaviorofthecave lakesduringthisextremeeventenabledustoformulate ideasaboutriver-groundwaterinteractionduringsuch majorhydrologicalevents.Theseobservationsprompted amoredetailedstudyoftherelationshipbetweentheriver andthekarstphreaticzone.Theobjectiveofthisstudyis thereforetodescribethehydraulicrelationshipbetweenthe BerounkaRiverandnearbycavelakesduringstableflow andfloodeventsandalsotodescribetheoriginofthe waterincavesanditspossiblerelationshiptotheriver water.Thesedataprovideimportantinformationaboutthe modernrateofspeleogeneticprocesses. R EGIONAL S ETTING T HE B OHEMIAN K ARST TheBohemianKarstisasmall,isolatedregionwitha totalareaofabout140km 2 consistingofSilurianto MiddleDevonianlimestonesthatwerefoldedandfaulted duringtheHercynianOrogeny.Theselimestonesare frequentlyfoldeddeepbelowthesurfaceandareinterbeddedwithnon-karsticrocks(Havl c ek,1989;Chlupa c etal.,2002).ThelimestonesareeithercoveredbyCretaceous platformsandstonesandmarlsorareoverlainbyfluvial riverterracesofTertiaryandQuaternaryage(Vc slova , 1980;KovandaandHerzogova ,1986;KuklaandLoz ek, 1993).Theareawherecavesandotherkarstfeaturescanbe directlyobservedisthereforemuchsmallerthanthefull 140km 2 area.WhilethenortheasternpartoftheBohemian Karstisflat,itssouthwesternpartishilly,withelevations between208and499ma.s.l.Deepvalleysoflocalstreams andespeciallyoftheBerounkaRivercutthroughthis southwesternpart.Relativeelevationdifferencesbetween therivervalleybottomanditssurroundinghillsreachmore than200m.Thepresentdayclimateoftheareaismoderate, withmeanannualprecipitationof493mm(measuredatthe Berounstation,1931–1960).Themeanannualairtemperatureisbetween8and9 u C. Theknowncavesaredevelopedmostlyinthe120to 300mthickLochkovianandespeciallythePragian limestone,ahigh-gradelimestonewithusuallymorethan 95percentCaCO 3 (BruthansandZeman,2001,2003).A totalof685mostlysmallcaveswithatotallengthof 23.1kmhavebeenmappedintheareaasofDecember31, 2010,accordingtotheCaveDatabaseoftheInstituteof GeologyASCR.Some59percentofthemwerefound whenexposedinlimestonequarries.Three-dimensional phreaticmazeswithmultipleloopsandirregularcrosssectionspredominateintheBohemianKarst.Thereare alsoseveralbathyphreaticcaves(BruthansandZeman, 2003).Theknowndepthofloopsinthebathyphreatic cavesexceeds100m;continuationsareeitherclosedby cavesedimentsortoodeepandnarrowtobeinvestigated bydiving(Podtrat 9 ova Cave;BruthansandZeman,2003). Theoriginofthecavesisnotyetfullyunderstood.Bosa k etal.(1993)believedthatthecavesmostlyoriginatedby mixing-corrosionofriverwaterwithwaterfromthe limestoneaquiferduringthelateCretaceousandPaleogene, whenthereliefwasrelativelyflat,withbroadrivervalleys. BruthansandZeman(2001)proposedtheideaofcave evolutionbydiffuserechargefromtheCretaceoussandstonesthatcoveredthewholeregioninthepastandby injectionoffloodwatersoftheBerounkaRiverandits Tertiaryprecursor,thePaleo-BerounkaRiver(Z a ketal., 2001b).Somecavesareclearlyhydrothermalinorigin(C lek etal.,1994;Bosa k,1998);buttherearenoindicationsfor hydrothermalorigininmostofthecavesinthearea. Themainpresent-dayregionalaquiferisdevelopedin thesamelithologicalsequenceasmostofthecaves (BruthansandZeman,2001).Basedonthermalmodeling, recentgroundwatercirculationinthemainaquiferoccurs insynclinesdowntodepthsofabout600mbelowthe surface(Z a ketal.,2001a).Sinksofsurfacestreams,stream caves,dolines,andothersurfacekarstfeaturescommonin classicalkarstareasaregenerallyabsentintheBohemian Karst(BruthansandZeman,2003). T HE B EROUNKA R IVERANDITS V ALLEY Thestudyarea(Fig.1)surroundsan8kmlongsection oftheBerounkaRivercanyonbetweenthetownsof BerounandKarls tejn.TheBerounkaRiveristheonly largestreaminthearea,withacatchmentareaof 8284.7km 2 ,meanflowrateof35.6m 3 s 2 1 ,calculated 10-yearfloodflow(Q 10 )of799m 3 s 2 1 ,andcalculated 100-yearfloodflow(Q 100 )of1560m 3 s 2 1 .Duringthe August2002flood,thepeakflowwas2170m 3 s 2 1 .(All dataarefromtheBeroungauge,locatedjustabovethe beginningofthekarstcanyon;datafromwww.chmi.cz,see alsoMoECR,2003).Duringnormalsummerwaterlevels, theriverelevationchangesfrom213to209ma.s.l.The meangradientofthewaterleveloftheBerounkaRiverin thestudiedsectionis0.64mperkm,locallyvarying R ESPONSEOFTHEKARSTPHREATICZONETOFLOODEVENTSINAMAJORRIVER (B OHEMIAN K ARST ,C ZECH R EPUBLIC ) ANDITSIMPLICATIONFOR CAVEGENESIS 66 N JournalofCaveandKarstStudies, April2012

PAGE 69

between0.4and1.1mperkm(Fig.2).Overbankflooding beginsapproximatelyataflowof450m 3 s 2 1 (i.e.,awaterlevelincreaseofmorethan3m).Localstreamsenteringthe BerounkaRiverintheBohemianKarsthavecatchments uptotensofkm 2 .OnlythelargerKac a k(Lode nice)stream hasalargercatchment,271.1km 2 ,andthemeanflowatits confluencewiththeBerounkaRiveris0.53m 3 s 2 1 .Mostof thecatchmentareaoftheKac a kstreamisoutsideofthe karstarea. Fromwrittenrecordsandfloodmarkspreservedinthe rivervalley,itcanbeconcludedthatfloodssimilartothe Q 100 peakfloworhigheroccurredseventimesduringthe last500years,in1598,1655,1675,1769,1784,1872,and 2002(Elleder,2004;Z a kandElleder,2007).Thehighest floodeverrecordedoccurredonMay25,1872,whenthe maximumwaterlevelinthevillageofSrbsko,inthemiddle ofthestudiedriversection,was7.7mabovetheusuallevel there,0.47mhigherthanduringthe2002flood(Z a kand Elleder,2007). TheAugust2002floodaffectedmostoftheCzech Republic.TheBerounkaRiverwaterlevelroseasaresult ofheavyrainsthatoccurredAugust6–13insouthwestern Bohemia.RiverpeakattheBeroungaugeoccurredat midnightonAugust13,withpeakflowof2170m 3 s 2 1 (www.chmi.cz).Themeanriverslopeduringthefloodwas similartothatofthenormalwaterstage,0.66mand0.64m perkm,respectively(datafromZ a kandElleder,2007;see Fig.2).Localvariationsinthegradientweresmaller duringthefloodcomparedtonormalflow. Figure1.ThestudyareaalongtheBerounkaRiverinthe BohemianKarst.Thefullnamesofthecavesaregiveninthe text.Thegeologyisbasedona1:25,000geologicalmap (Havl c ek,1989). Figure2.ProfileoftheBerounkaRiver(Z a kandElleder,2007).Trianglesrepresenttheelevationsofthecavelakes.The lowerlineandsolidtrianglesarefornormalflow(100cmontheBeroungauge),andtheupperlineandopentrianglesarefor thepeakofthe2002flood.Thegradients( % ormperkm)areindicatedforlocalstretchesoftheriver. H.V YSOKA ,J.B RUTHANS ,K.Z A K ,J.M LS JournalofCaveandKarstStudies, April2012 N 67

PAGE 70

TheevolutionofthestudiedsectionoftheBerounka RivervalleyduringtheTertiaryandQuaternarywasquite complex.StartingintheOligoceneandearlyMiocene,the paleo-BerounkaRiverformedawidevalleywithbottom locatedabout70mabovethepresentriverbed(Kuklaand Loz ek,1993;Z a ketal.,2001b).Thesamewidevalleywas usedagainduringtheearlyPleistocene.Duringthe Pleistocene,after780kaBP,asindicatedbypaleomagneticandpaleontologicalstudies(Hora c ekandLoz ek, 1988;Koc ,1991;Kovanda,1991),theBerounkaRiver startedtocutanarrowcanyon,inwhichindividualriver terraces,correspondingtomiddlePleistoceneclimatic changes,arepreservedlocally.TheareaoftheBohemian KarstwasneverglaciatedduringtheQuaternary.During thelastglacialperiod(Weichselian),thedeepestlevelof theriverbottomwasseveralmetersbelowthepresentbed. DuringthelateWeichselianorearlyHolocene,thebottom ofthevalleywasfilledinwithgravelandsand,forminga flatfloodplain.Thesecoarse-grainedalluvialsediments are8.0to13.5mthick,basedonboreholedata(Vc slova , 1980),andareusuallycoveredbya0.5to2.5mthicklayer oflateHolocenefine-grainedfluvialsediments(Z a ketal., 2010). Thehydraulicconductivityofthecoarsealluvium rangesbetween10 2 4 and10 2 2 ms 2 1 ,basedonpumping testsconductedoverthearea(Vc slova ,1980).Thevertical hydraulicconductivityofthelateHolocenefine-grained fluvialcoverofthefloodplainvariesbetween10 2 5 and 10 2 4 ms 2 1 ,basedonfieldinfiltrationtests.Thehydraulic conductivityofthelimestoneaquifer(mainlyfractureand karsticporosity)isbetween10 2 8 and10 2 4 ms 2 1 ,basedon pumpingtestsinthearea(Vc slova ,1980).Thestorativity ofthelimestoneaquifer,derivedfromintegrationofthe springyieldtogetherwithmonitoringofthewaterlevel recessionattheboreholes,is3to7 % (Bruthansand Zeman,2000). S TUDIED C AVES Sixcavesthatcontainsmalllakesandarelocatedon bothsidesoftheriverwerechosenforregularmonitoring (Table1,Fig.1).Fiveofthemaresituatedrelativelyclose totheBerounkaRiver:MenglerovaCave(abbreviated MEN),whichisconnectedwiththeEmenta lCave(EME) byanopenconduit(sump),DynamitkaCave(DYN), Podtrat ova Cave(POD),Toma s kovaChasm(TOM),and Tet nsky Vy ve rCave(TET).Thesecavesoccurbetween40 and190mfromtheriverchannel.TheNova naDamilu Cave(NOV),situatedapproximately1kmawayfromthe riveratamuchhigherelevation,withnoriverinfluencebut withhighlyvariablewaterlevel,wasalsoincluded.Since thestudiedcavesarelocatedwithinthesteepslopesofthe canyon,theunsaturatedzoneabovethecavelakesis severaltensofmetersthick.Undernormalflowconditions, theriverchannelisnotdirectlyconnectedwithanyof studiedcavelakes.Thealluvialsedimentsblockdirect, rapidinfiltrationofriverwaterintothelimestoneaquifer. Underthefloodregime,thewaterlevelintheriverreaches solutionalopeningsinthelimestonesidesofthecanyon Table1.Parametersofthecavelakes. CaveAbbreviation Cavelength (m) Cavelake abovethe Berounka river(m) Lakedepth (m) Lakesurface (m 2 ) Distance fromriver a (m) Distance fromriver b (m) Landuse abovecave PodtratovaPOD180 6 0.367185520unused; bushes TomaskovaTOM80 6 0.325170120abandoned quarry; bushes DynamitkaDYN180 6 0.31.512190140abandoned quarry; forest Menglerova/ Emental MEN/EME15/1884 6 0.3614/50400unused; bushesagricultural land TetinskaTET8.5 + 1.90,50–8500unused; bushesbuildings NovaNOV156 + 700–150–7512001200abandoned quarry; bushes a Nominalstage. b Floodculmination2002. R ESPONSEOFTHEKARSTPHREATICZONETOFLOODEVENTSINAMAJORRIVER (B OHEMIAN K ARST ,C ZECH R EPUBLIC ) ANDITSIMPLICATIONFOR CAVEGENESIS 68 N JournalofCaveandKarstStudies, April2012

PAGE 71

(mostlytheentrancesofthecaves).IntheMEN-EMEcave system,sedimentaryevidenceshowsthatfloodwaterscan rapidlyreachevenverydistantpartsofthecave,andisa perfectexampleofaflood-flowlabyrinth(sensuPalmer, 1991). Quarriesclosetothestudiedcaveshavequarrybottoms abovethehighestgroundwaterlevelandthehighest floods.Quarrydebrisdoesnotblocktheartificialentrances anywhere.Abovetheactivephreaticandepiphreaticzones, thestudiedcaveshavedry,nowinactivesectionsthat extenduptoseveraltensofmetersabovethegroundwater level.Thesedrysectionswerelocallypartlyquarriedout, butdidnotimpactondeepercavesections.Therefore,the influenceofquarryingonthehydraulicregimeofstudied cavescanbeconsideredtobenegligible. M ETHODS ThemaximumwaterleveloftheAugust2002flood inthevalleyandinthecaveswasmarkedforlater measurements.Aftertheflood,plasticwater-gauginglaths wereplacedandfixedbyscrewsinthesixcaves,aswellas onthebankoftheBerounkaRiverclosetothecaves.The altitudesandpositionsofthelathsoutsidethecavesand caveentrancesweremeasuredbyatotalstationwith precisionof 6 5mm(SokkiaandTopcon).Thepeakwater levelsofthe2002floodatseveralprofileswithinthestudied valleyandtheelevationsofhistoricalfloodmarkswere measuredinthesameway.ThedataofZ a kandElleder (2007)wereusedforcomparison.Theelevationsofthe lathsinsidecavesweremeasuredbyapreciseinclinometer andalaserrangefinder(DistoLite5,LeicaCo.). Threecavelakes(inMEN,DYN,andPODCaves) wereequippedwithpressureandtemperaturesensors (Lucas)connectedwithdataloggersandautomatically measuringthewaterlevelandtemperatureevery30minutes withanaccuracyof 6 1cmand 6 0.03 u C,respectively (LGR,GeomonCo.).Thesubmergedpressuresensorsare connectedtotheatmospherebyacapillarytubesto compensateforairpressurechanges. Theconductivity,temperature,andpHweremeasured inthecavelakesandriverwaterusingportabledevices (Cond340i,pH330i,WTWCo.)onamonthlybasis.The cavelakesandtheBerounkaRiverweresampledmonthly forchloride,aconservativetracer.Samplesformajor chemicalanalysisweretakenduringhighflow(April2003) andlowflow(September2003).Samplesfor 13 Candmajor chemistryweretakeninOctober2007. WaterchemistrywasanalyzedbyFAAS,HPLC,and titrationinthelaboratoriesoftheCzechGeological Survey.Samplesforchloridecontentwereanalyzedby argentometrictitration.Samplesforthe 13 Cdetermination Figure3.Themodelofthehydraulicrelationshipbetweentheriverandacavelake.SeeEquations36.A.Thegeometryof themodel.B.RelationshipofQ 1 andQ 2 totheaxesinFigure6.C.AnexampleofapossibletimeevolutionofQ 1 andQ 2 H.V YSOKA ,J.B RUTHANS ,K.Z A K ,J.M LS JournalofCaveandKarstStudies, April2012 N 69

PAGE 72

ofHCO { 3 werecollectedinvapor-tightcontainerswith zeroheadspace,andthebicarbonatewasimmediately precipitatedasBaCO 3 inthelaboratory.Fortheisotope measurement,theBaCO 3 wasconvertedtoCO 2 bytheusual reactionwith100%H 3 PO 4 .The 13 Cmeasurementswere performedonaFinniganMat251MassSpectrometerinthe laboratoriesoftheCzechGeologicalSurveywithan analyticalerrorof 6 0.1 % .Theresultsareexpressedinthe usualdeltanotationagainsttheinternationalPDBstandard. Thesaturationindexeswithrespecttocalcitewere calculatedfromthechemicalanalysesdataandfieldpH measurementsusingthePHREEQC2.14code(Parkhust andAppelo,1999). Communicationbetweentheriverandthelimestone cavesthroughthealluviumwasmodeledasnonstationary 1Dflowinaverticalplanenormaltotheriverflowand assumedanunconfinedaquifer.Themodelwasgoverned bythenonlinearparabolicpartialdifferentialequation S L u L t ~ L L x Ku L u L x 1 where x isthedistancefromtheriver, t istimeand u isthe hydraulichead.Themodelcontainstwoparameters, hydraulicconductivity K andthestorativity S ofthe aquifer.TheRothemethodwasappliedtodiscretizethe problemintime,andtheGalerkinmethodwithfinite elementswasutilizedwhensolvingtheresultingelliptic boundaryvalueproblem.Theoutputdataarethe hydraulichead u anddischarge q givenby q ~{ Ku L u L x : 2 Themodelandthemeasureddatasetswereusedto solvetheinverseproblemofdeterminingthemodel parametersatthethreelocations.Thefirstresultsmade itobviousthattheconnectingpathsbetweentheriverand cavelakesareunderconfinedaquiferconditions.This allowedtheauthorstointroducetransmissivity T andto simplifythemodelinthelinearform S L u L t ~ T L 2 u L x 2 3 Equation(3)wasusedtosimulatethehydraulichead propagationbetweentheriverandthecavelakes.The lowvaluesofthestorativitymadeitpossibletoassume S 5 0andtofurthersimplifythemodel,illustratedin Figure3,to Q 1 ( t ) ~ W H L ( t ) { H R ( t ) R 4 H L ( t ) ~ H L t { D t z D t Q 2 t { D t { Q 1 t { D t A 5 ormoresimply Q 2 t ~ A H L t z D t { H L t D t z Q 1 t 6 where R ~ L T : 7 H R (t)istheriverlevel, H L (t)isthelakelevel, Q 1 (t)isthe flowbetweenthecavelakeandtheriver, Q 2 (t)istheflow betweenthecavelakeandthelimestoneaquifer. W isthe widthoftheaquiferparalleltotheriver(m), A istheareaof thecavelake(includingunknowncavelakesinthevicinity ofthecave), R isthehydraulicresistivityoftheconnecting pathbetweentheriverandthecavelake, T isthe transmissivityoftheflowpath(m 2 s 2 1 ),and L isthedistance betweenthecavelakeandtheriveralongtheflowpath. R and A areconsideredtobeconstantintime,whichmaybea considerablesimplificationofreality(seebelowinthe discussionofthesimulation). Q 1 (t), Q 2 (t), R ,and A areunknown,butplottingthemeasuredvaluesof H L (t + D t) 2 H L (t)/ D tand H L (t) 2 H R (t)foreachcavelakeinFigure3 enablesustodirectlydetermineif Q 1 (t)and Q 2 (t)are positiveornegativeandwhichoneisincreasingor decreasingatagiventimeinthecourseoftheflood. R ESULTSAND D ISCUSSION R ELATIONSHIP B ETWEENTHE W ATER L EVELINTHE R IVERANDINTHE C AVE L AKES Atsteadystate,thewaterlevelsinthecavelakesareat thesameelevationasthewaterlevelattheclosestpointin theriverwithinameasurementerrorof 6 0.3m.Theonly exceptionsaretheTETCave,withanaveragewaterlevel of + 1.9mabovetheriver,andtheNOVCave,whichis about1kmawayfromtheriver. Atotaloftwelvewaterlevelpeakswithamplitudesof upto1.5mweredocumentedbetweenJanuaryandJuly 2004(Fig.4).Thewateroscillationsinallthecaveswere verysimilar(thePearsoncorrelationcoefficientamongthe cavelakeswasbetween0.977and0.990).Theriverwater didnotdirectlyenterthecavesduringtheperiodof automaticlogging.Theriverandthecavelakeswerethus hydraulicallyconnectedonlyviathesaturatedzoneofthe Quaternarygravelsintheriveralluvium. TheBeroungaugingstationusedformonitoringthe waterlevelintheriverissituated5.5to7.0kmupstream fromcavelakesPOD,DYNandMENincludedin Figure4.Propagationofthefloodpeaksdownthe BerounkaRiveroccurredataveragespeedsof3.5to 7.0kmh 2 1 ,basedonnineflowpeaksin2004.Thedelay betweengaugingstationandtheriverinfrontofcavesis thusinorderof1to2hours,whichcanbeneglectedsince individualpeakslasttensofhoursandmore.Therewasa highcorrelationbetweenthelevelsoftheriverandofthe R ESPONSEOFTHEKARSTPHREATICZONETOFLOODEVENTSINAMAJORRIVER (B OHEMIAN K ARST ,C ZECH R EPUBLIC ) ANDITSIMPLICATIONFOR CAVEGENESIS 70 N JournalofCaveandKarstStudies, April2012

PAGE 73

Figure5.DetailfromFigure4ofMarchandearlyApril2004.SeethecaptionforFigure4forexplanation. Figure4.WaterlevelintheriverattheBeroungaugeandwaterlevelsinsomecavelakes.TheyieldoftheKodaSpring (Fig.1;rightaxis)isshownforcomparison.Notethattheverticalpositionofalaketracemaybeoffbyupto0.3m.PeaksA andBareduetowintericejamsdownstreamfromthecavesthataffectedtherivernearthecavesbutdidnotaffecttheriverat thegauge.Thetimescaleisday.month.yearafter2000. H.V YSOKA ,J.B RUTHANS ,K.Z A K ,J.M LS JournalofCaveandKarstStudies, April2012 N 71

PAGE 74

cavelakes.Oncetheriverlevelstartedtorise,thelake levelsroseaswell,butslightlymoreslowly.Duringallthe events,thefastestreactionwasobservedinthePODCave, andtheslowestandleastprominentincreasewasinthe DYNCave(Fig.5).Thisprobablyreflectsthelonger distancebetweentheriverandtheDYNCave(Table1). Thecavelakelevelsweremostlystillrisingwhentheriver levelstartedtofall.Lakelevelsstartedtodeclinewhenthe riverlevelbecamelowerthanthelakelevels(Fig.5).The lakelevels’declinewasveryslow,takingseveralweeksto severalmonthstoreturntotheriverlevel.Shortperiods withtheriverlevelabovethatofthecavelakes,compared torelativelylongperiodsoftheoppositerelationshipshow thatriverwatertendstoflowintothelimestoneaquifer onlyduringfloodpeaks.Duringmostoftime,waterdrains fromthelimestoneaquiferintotheriver. Duringthewinter,twowaterlevelpeakswereobserved inthecavelakesthatwerenotconnectedtopeaksinthe riverlevel(periodsdesignatedAandBinFig.4).Nothing similartothesepeakswasobservedinthewaterlevelsof localwellsorspringsorinstreamdischarges.These oscillationsofthecavelakeswaterlevelswereclearly causedbyaslush-icebarrierformationontheBerounka RiverabovethetownofKarls tejn,immediatelybelowthe studiedriversection.Icejamsformrepeatedlyatthissite, sincethefloatingslushiceaccumulateshereattheupper edgeofalreadyfullyfrozen,almoststagnantwaterabove theKarls tejnweir.Thankstothemodestrivergradient,the backfloodingaffectstheriverleveluptoseveralkilometers abovetheicejamandconsequentlyalsoaffectsthe monitoredcavelakes,butitdidnotreachtheBeroun gaugingstationsituatedseveralkmfartherupstream.The existenceofthetemporaryicejamswaslaterconfirmedby iceaccumulationshighonthebanksoftheriverchannel. Unfortunately,nolargefloodwithwaterfillingthewhole floodplainandflowingdirectlyintothecaveopenings occurredduringtheperiodofautomaticmonitoringofthe cavelakelevels. Thehydraulicrelationshipbetweentheriverandthecave lakeslevelsispresentedinFigure6.Timeintervalsaffected byicejamswerenotusedinthegraphsandcalculations. Boththeriverandthelakelevelweremeasuredvery preciselyovertime( 6 1cm),butthereisconsiderable uncertaintyintherelativepositionsofthesetwodatasets ( 6 0.3m).Therefore,inpartsA–CofFigure6,thereal positionofthewholedatasetmaybeshiftedeithertotheleft ortotherightalongthehorizontalaxis.Inallthemonitored caves,thetemporalchangeinthecavelevelisclearly controlledbyadifferencebetweenthelevelintheriverand thatinthecave.Thereisalinearrelationshipbetweenthese twoparameterswithcoefficientsofdeterminationbetween 0.55and0.78.Theindividualmeasuringpointsdonot followtheregressionlineexactly.Instead,thereissome tendencyforhysteresis(mostpronouncedinthePOD Cave).Temporaltracksofpointsintheupperleft-handpart ofthegraphs,correspondingtotherisinglimbsofcavelake hydrographs,followpathsinacounterclockwisedirection (seethearrows).Thetrackscanbeinterpretedbymeansof Figure3,partsBandC,andFigure6D.Whentheriverlevel risesabovethelakelevel( H L (t) 2 H R (t) 0anddecreasing), Q 1 bringswatertowardsthecavelakeproportionallytothe difference H L (t) 2 H R (t). Q 2 ispositiveandthewatertablein thecaverises.When H L (t) 2 H R (t)ismostnegative, Q 2 abruptlyfallsto0oreventonegativevalues(seeFig.3Afor themeaningofnegativeandpositive Q 1 and Q 2 values).Then theabsolutevalueofQ 1 decreasesproportionallytothe decreaseintheabsolutevalueof H L (t) 2 H R (t). R EACTIONOF C AVE L AKESTOTHE E XTREME F LOODIN A UGUST 2002 Thecavelakeswerevisitedtwiceduringthismajor floodevent.ThefirstvisittookplaceonAugust14about 12hoursaftertheriverpeak,whentheriverlevelatthe Beroungaugewas710cm,followingapeakof796cm.The secondvisitwasonAugust16,whentheriverlevelhad fallento400cmontheBeroungauge. ThemaincauseofthecatastrophicAugust2002flood wasintenseprecipitation,especiallyinthesouthernand southwesternpartsoftheCzechRepublic,wherethetotal precipitationintheperiodbetweenAugust6andAugust15 exceeded300mminthemountainregions.Precipitation occurredintwoseparateperiods,betweenAugust6and7, andbetweenAugust11and13,whenalargeramountfell.In theBohemianKarstitself,theprecipitationfromAugust6 toAugust13reached130to160mm,approximatelytwice thenormaltotalprecipitationforAugust.Asaconsequence, thegroundwaterlevelrosequickly,andintensedripwater fromtheunsaturatedzonewasencounteredinsomecaves duringtheAugust14visit. Figure2showstheresponseofthecavelakestothis majorflood.Theirbehaviorcanbesplitintothefollowing groups:A)TheBerounkaRiverdirectlyfloodedthe entrancepassagesoftheTETandMENCaves.Afterthe riverpeaked,thewaterlevelinthecavesrapidlydecreased tothelevelofthebottomofthecaveentrance.Thenthe waterlevelinbothcavesdecreasedveryslowlyduetothe lowpermeabilityofthesedimentaryfillbetweenthecave lakesandtheriverandalsobecauseoftheintenseinflowof groundwaterfromthelimestoneaquiferintothecaves.In theTETCave,thewatertablefellslowlyforaperiodof 1.5years;B)IntheTOMandDYNCaves,situated120to 140mawayfromtheriver,notracesofdirectriverwater intrusionwerefound.Thisisbasedonconductivity(Fig.7) andtemperaturemeasurements.Nomuddywaterwas observedinthecavelakes.IntheTOMcave,thewater levelrose3.5m,asobservedduringthevisit12hoursafter thepeakoftheflood.IntheDYNcave,thewaterrose 3.6m,andthewaterlevelhaddroppedonly0.2mbelow themaximumlevelrecordedinthiscavebytheAugust16 visit,60hoursafterthepeak;C)ThewaterofthePOD Cavelakewascomposedofamixtureofriverwaterwith groundwater(temperature14.2 u Candhighturbidity,but R ESPONSEOFTHEKARSTPHREATICZONETOFLOODEVENTSINAMAJORRIVER (B OHEMIAN K ARST ,C ZECH R EPUBLIC ) ANDITSIMPLICATIONFOR CAVEGENESIS 72 N JournalofCaveandKarstStudies, April2012

PAGE 75

higherconductivitythanriverwater;Fig.7).Waterinthe cavereachedthesamelevelasintheriver.Thedistanceof thecavelakefromtheriverwasonly20matthepeakof theflood. T HE M ODELOF H YDRAULIC H EAD P ROPAGATIONAND F LOW B ETWEENTHE R IVERAND C AVE L AKES Equation(3)governsthehydraulic-headpropagation betweentheriverandthecavelakes.Solutionoftheinverse problemsforthetransmissivityofthealluviumbetween 10 2 3 and10 2 2 m 2 s 2 1 demonstratedthatthestorativityis # 10 2 3 fortheconnectingpathbetweentheriverandcaves DYN,MEN,andPOD.Thelowvaluesofthestorativity madeitpossibletofurthersimplifytheproblemto Equations(4)–(6),whichprovidethemodelfortheflow showninFigure3. TherelationshipbetweenthedatainFigures6A–Cand thismodelisshowninFigure6D.Theslopeindicatedby angle a isinverselyproportionaltotheproduct R 3 A .The PODCavehasthehighestslopeandthusthelowest R 3 A value,whileDYNhasthelowestslopeandthusthehighest R 3 A value.Thattheinterceptoftheregressionlineis slightlygreaterthanzeroforthePODandMENcaves indicatespermanentinflowintothecavelakethatis drainedintotheriver( Q 2 ). TheDYNCavelakeleveloscillationduringthe2002 floodwassimulatedusingtheslope 2 0.0091ofthelinear regressionbetweenthedifferenceintheDYNCavelakeand theriverlevelandthechangeinthecave-lakelevelin30min timeincrements(Fig.6A).ThesimulatedlevelintheDYN Cavefitsverywellwiththemeasuredlevelinthecave (Fig.8).Thisshowsthatfloodwaterusedthesamepath Figure6.Therelationshipduringthemonitoredperiodbetweenthedifferencebetweenthecave-lakeandriverlevelHL(t) HR(t)andthechangeinlakelevelHL(t)HL(t1)betweenobservations.Pointsareconnectedintimeorder.Boldarrows indicatetrajectoriesduringpeaksoffloods.Notethatthepositionsofthedataforagivencaveinthehorizontaldirectionmay beconsistentlyoffbyupto0.3m.A.DYNcave.B.PODcave.C.MENcave.D.Illustrationoftherelationoftheseplotsto themodel. H.V YSOKA ,J.B RUTHANS ,K.Z A K ,J.M LS JournalofCaveandKarstStudies, April2012 N 73

PAGE 76

betweentheriverandthecavelakewiththesame parametersduringthemajorfloodin2002asduringthe muchlowerwaterstageswhenthelevelwaslogged.R 3 A thusremainedconstantevenduringthisextremeflood. Noadditionalconduitwasactivatedduringthefloodin theDYNCave. Theslopeofthelinearregression(Fig.3B,C,6) canbepotentiallyusedforsimulationofthelakelevel inotherobjectsiftightcorrelationexistsbetween ( H L (t + D t) 2 H L (t))/ D t and H L (t) 2 H R (t).Theslopeofthe regressionlinecanbeusedforsimulationofthewatertable evenwhentherelativepositionofthelevelintheriver(or, Figure8.ThefloodontheBerounkaRiverinAugust2002,withsomemeasuredlevelsofthecavelakes.Notethatthepeak levelofthelakeinDYNisfromtracesleftinthecave,andthetimingofthepeakisuncertain.Theelevationsofthecavelake observationsrelativetotheriverareuncertainbyupto0.3m.Themodellineisasimulationbasedonthemodeldescribedin thetextandtheDYNcaveparametersderivedfromFigure6.Thehorizontalaxislabelsareday,month,andyear. Figure7.Conductivityandchloridecontentsinriverwaterandselectedcavelakesfollowingthepeakofthe2002flood. PointsarelabeledbythedayinAugust2002. R ESPONSEOFTHEKARSTPHREATICZONETOFLOODEVENTSINAMAJORRIVER (B OHEMIAN K ARST ,C ZECH R EPUBLIC ) ANDITSIMPLICATIONFOR CAVEGENESIS 74 N JournalofCaveandKarstStudies, April2012

PAGE 77

ingeneral,anydrivingobjectingeneral)andthelake(or anydependentobjectsuchasawell)isnotknownexactly. C AVE L AKE W ATER T EMPERATURE Thewatertemperatureinthecavelakes(Fig.9)is stronglyaffectedbyexposureoftheterrainabovethecave toradiationfromthesun.WhiletheTOMCave,whichis shieldedfromdirectsolarradiationbythesubvertical north-facingsideoftherivercanyon,hasanannual averagetemperatureof8.4 u C,theMENCave,situated underasouth-facingslopejust400maway,hasanannual averagetemperatureof11.1 u C. ThelaketemperatureinthePODCaveisstable(9.5 6 0.06 u C),exceptforextremefloodeventslargerthanthose duringtheperiodofdatacollection.TheDYNCave exhibitsaslighttemperatureoscillationovertherange9.0– 9.3 u C,andthewatertemperatureisdirectlyproportional tothelakewaterlevel( r of0.62).Thiscanbeinterpretedas ariseofwaterwithslightlyhighertemperaturefroma deeperpartofthecavelaketowardsthewatertableduring arisinglakelevel.TheTOMandNOVCavelakesalso exhibitonlyverylimitedtemperaturevariations,basedon manualmeasurements(lessthen 6 0.2 u C). IntheMENCave,thewatertemperaturerangedfrom 10.6to11.7 u C.Thetemperaturewashighestinlate NovemberandlowestinJuneduringthemonitoring period,whichistheoppositetotheseasonaltrendofthe groundsurface.Itstemperaturegraphconsistsofasmooth sinusoidalannualoscillationwithshort-timedeviations correlatedwithlake-levelchanges.Thewatertemperature decreasesduringbriefriverrisesevenwhentheriverwater iswarmerthanthewaterofthecavelake.TheMENCave lake’swater-temperatureoscillationsaretypicalforcombinedheattransport(conductiontogetherwithconvection),basedonthestudyofBundschuh(1997).Sinusoidal oscillationofthelakewatertemperatureisgeneratedby annualchangesinthesolarradiationthatdirectlyaffectthe soiltemperature.Thisvariationpropagateswithadelay,to Figure9.Measuredtemperaturesofthecavelakes(leftaxis)andriverandoutdoorair(dailyaverage;rightaxis).Thewater levelinthePODcavelakefromFigure4issuperimposed(noscale).Thehorizontalscaleisthemonthandyearduringthe periodofdetaileddatalogging. H.V YSOKA ,J.B RUTHANS ,K.Z A K ,J.M LS JournalofCaveandKarstStudies, April2012 N 75

PAGE 78

depthbyheatconductionviarocks.Short-livedpeakson thesinusoidaltrendarearesultofintrusionofwaterwith lowertemperature,mostprobablyfromtheQuaternary gravelsofthefloodplain. C HEMICAL C OMPOSITIONOFTHE C AVE L AKE W ATERAND ITS O RIGIN Thewaterinthecavesstudiedcanbederivedfrom twopossiblesources,theBerounkaRiverandlocal groundwater.About500pairsofconductivityand chloride-contentvaluesweremeasuredforsprings, streams,andcavelakesand poolsintheBohemianKarst inthe2000–2004period.ThewateroftheBerounka Riverhaslowconductivityandthuslowtotaldissolved solids(TDS),thankstothelocationofmostofits catchmentonsilicaterocks,abovethestudiedkarstarea. Thechloridecontentoftheriverwateriselevated,thanks tomultiplesourcesofpollution,bothurbanand agricultural.Theconductivity-chlorideplotofthewater oftheBerounkaRiverdiffersfromthoseofallthe groundwaterandcavelakewatertypesoftheBohemian Karst(Fig.10). Springsandwellslocatedbothinthelimestoneandin thesurroundingnon-karsticrocksshowconsiderably higherconductivitythantheriverwaterthankstohigher TDS.Thechloridecontentvariesfrom6tomorethan 50mgL 2 1 ,dependingonthedegreeofpollutioninthe catchmentofaparticularspring.Thecavelakeofthe C er nkaCave,situatedunderanopenlimestonequarry 3kmawayfromtheriver(Fig.1),exhibitsachloride contentaslowas2mgL 2 1 ,thankstotheverylow evapotranspirationenrichmentofchlorideionsinwater infiltratingintolimestonefractureswithlackofsoilcover removedbyquarrying.Cavelakesandpoolsthatretain waterfromtheunsaturatedzonecouldbesplitintotwo groups,basedontheirobservedchemistry. Someofthelakesandpoolsfallinsidethefieldtypicalfor thekarstsprings(Fig.10)andtheirchemicalcompositionis Figure10.Plotofconductivityversuschloridecontentofwatersamplesanalyzedduringthestudy.Theareaoccupiedbythe riversamplesisclearlydistinctfromcavelakes,seepage,andotherundergroundwater.Degassingandcalciteprecipitation moveundergroundwatershorizontallytotheleftasshown,becausechlorideisconserved.Thespringdata( + )arefromthe BohemianKarst,measuredin20002004.Thechloridebackgroundindicatedistheapproximateminimumofthosesamples; near-surfaceundergroundwatercanbelower.TheCerinkaquarryis2kmnorthofthestudyarea(Fig.1). R ESPONSEOFTHEKARSTPHREATICZONETOFLOODEVENTSINAMAJORRIVER (B OHEMIAN K ARST ,C ZECH R EPUBLIC ) ANDITSIMPLICATIONFOR CAVEGENESIS 76 N JournalofCaveandKarstStudies, April2012

PAGE 79

alsothesame(NOVCave,TETCave,partofEMECave. Theselakesarerechargedbylocalgroundwaterwithno subsequentchangesinthewaterchemistry. Somelakesandpoolsexhibitlowconductivityandlow chloridecontent(theTOM,DYNandPODCavesand someoftheEMECavepools).Ontheconductivity-chloride plot,theyfallneitherinthefieldofkarstspringsnorinthe areaofBerounkaRiverwater(Fig.10).Aschlorideisa conservativetracer,thesewaterscannotbederivedfrom riverwater,whichhasamuchhigherchloridecontent.The calciumandalkalinitycontentsinsomeofthesewatersare uptotwoorthreetimeslowerthanthoseofthekarst springs.Othercomponentshavesimilarconcentrationsto thosemeasuredinthekarstsprings.Thesecompositionscan beexplainedbyprolongedequilibrationofthewaterwitha caveatmospherecontaininglowCO 2 partialpressure(0.04 to0.18vol.%CO 2 inthestudiedcaves). Long-termstagnationofwaterinlakesandpoolsand highwaterpHareinagreementwiththisexplanation,as aretheirhigh d 13 Cvalues(Fig.11).TheTOMCavelake showsthemostsignificantlossofcalciumandbicarbonate ionsbyCO 2 degassingandcalciteprecipitation.Itcontains only66–70mgL 2 1 calciumcomparedto140–200mgL 2 1 typicalforthekarstspringsofthearea.IntheTOMand PODCaves,thecontentofbothcomponentsvariesslightly overtime,probablyduetovariationsintheintensityof flowthroughthecavelakeandmixingofstagnantlake waterwithwaterfromtheunsaturatedorsaturatedzones. Generally,thelowestTDSandhighestpHvalueswere detectedduringthewinterof2003–2004,afteramorethan year-longrecessioningroundwaterlevels.Theverylow chloridecontentintheTOMCaveshowsthatthecavelake ispredominantlysuppliedfromdirectinfiltrationinthe quarryabovethecave. Westudiedthetemporalbehaviorofchloridetodiscover whetherthecavelakesarepartlyrechargedbyriverwater (Fig12).Thefigureclearlydemonstratesthatthetemporal trendsinriverwaterandcavelakesaredifferent.Duringthe recession(March2003toOctober2003),thechloride contentgenerallyroseintheriverwater,whereasitwas quitestableinthecavelakes.Itdecreasedslowlyinthe MENCavelake,probablyasaconsequenceofdilutionof theriverwaterinjectedduringthe2002floodbylocal groundwater.Inflowofriverwaterwasthusnotobservedin anyofthestudiedcavesduringlowandnormalwaterstages. Thecavelakeswereslightlyoversaturatedwithrespect tocalcite(SI0.2to0.7).Riverwaterwasundersaturated withrespecttocalciteduringhighwaterstages(SI 2 0.5) butoversaturatedduringlowandnormalwaterstages(SI 1.0to1.3). I NTERACTION B ETWEENTHE R IVERANDTHE L IMESTONE A QUIFER :AM ODEL Basedonobservationsofthewaterlevelinwellsandcaves locatedfartherawayfromtheriver,theminimumhydraulic gradientinthelimestoneaquiferisbetween1and2%,and locallyhigher.Theriverslopeissignificantlyless(0.064%,on anaverage,seeFigure2).Thereforethegroundwatertendsto Figure11.Bicarbonateand d 13 Cvaluesforsomesamplesanalyzedinthisstudy.Bohemiankarstspringsfallinthearea outlined.Asanexampleoftheeffectofdegassing,valuesfortheCisarskaRokleSpring(Fig.1)areshownatthespringand after1kmofflowonthesurfacetowardtheriver(*). H.V YSOKA ,J.B RUTHANS ,K.Z A K ,J.M LS JournalofCaveandKarstStudies, April2012 N 77

PAGE 80

dischargeintotheriveratsteadystate.Riverwaterentersthe limestoneaquiferonlytemporarily,duringfloodeventsor duringwintericejams,whenthehighwaterlevelintheriver canyoninjectsriverwaterintothelimestoneaquifer.During majorfloods,thewaterlevelinthelimestoneaquiferisseveral metersbelowthewaterlevelintheriver,thusenablingvery effectiveinjectionoffloodwatersintothekarstporosityupto adistanceofafewhundredmetersawayfromtheriver(e.g., theEMECave).Afterfloods,waterreturnsbacktotheriver bythesamepath. ObservationsofcavelakesduringtheAugust2002 floodclearlydemonstratedtheimportanceofopencave entrancesforfloodingbyriverwater.Caveswithopen entrancesatelevationsaccessibletofloodintrusion wereextensivelyfloodedbyriverwater.Cavesseparated fromtheriverbyaQuaternaryalluviumwereflooded byelevatedgroundwaterinsteadofriverwater.Thewater levelinsuchcavesrosetoonlyabouthalfoftheriverÂ’s peaklevel.Riverwaterpenetratedonlyafewtensofmeters intothelimestoneaquiferinthesecases. Undercurrentconditions,massiveflood-flowinjection onlyoccursincaves,suchasMEM,EME,andTET,with exposedopeningsbelowthelevelofriverfloods,which,asin 2002,canbeupto7mabovethenormalriverstage.Incaves separatedfromtheriverbyQuaternaryalluvium,flood-flow injectionisstronglyreducedorabsentatthepresenttime.In thosecaves,flood-flowinjectionwasaveryimportantprocess duringsomeperiodswithinthelasttwoglacialperiods,when therockyfacesoftherivercanyonwerenotcoveredbyfinegrainedsedimentsandriverwaterwaseasilyinjectedinto fracturesandcaveopenings.Atpresent,theflood-flow injectionisalsoreducedbyanthropogenicfactors,especially the1to2mthicklateHolocenefine-grainedfluvialsediments coveringthefloodplainsurface,largelytheresultofenhanced erosionduetodeforestationandextensionofagriculturein therivercatchmentsincethethirteenthcentury. Theshapeofverticalsectionofthecavesisalsoan importantfactorfortheextentofriver-waterpenetration intothelimestoneaquiferduringaflood(Fig.13).In highlypermeableshallowcavepassages,wherepartofthe passagesissituatedintheepiphreatic(periodicallyflooded) zone,theriverwatercanpenetrateseveralhundredmeters intothelimestoneaquifer,asthereisenoughroomfor incomingriverwater.(EMECaveisanexample.)Inhigh volumephreaticloopsconnectedtoonlysmallspacesin theepiphreaticzone,therisingwaterfillslessvolume. Consequently,riverwatercannotpenetratefartherintothe limestoneaquifer(TOMandPODcaves). Figure12.Chloridecontentintheriverandmonitoredcavelakes.Thedischarge(notlevel)oftheriverattheBeroungaugeis shownforcomparison,asistheyieldoftheKodaSpring(Fig.1).Thetimescaleisday.month.yearafter2000.Notethatthe periodcoveredisdifferentfromthatofFigure4. R ESPONSEOFTHEKARSTPHREATICZONETOFLOODEVENTSINAMAJORRIVER (B OHEMIAN K ARST ,C ZECH R EPUBLIC ) ANDITSIMPLICATIONFOR CAVEGENESIS 78 N JournalofCaveandKarstStudies, April2012

PAGE 81

Onenewresultofthisstudywastheobservationofwater levelincreasesintheriver(andinthecavelakes)relatedto icejamsblockingtheriverandthuselevatingtheriverlevel duringthewinter.Theseicejamsareoftwotypes,formedby eithertheaccumulationofmovingslushiceduringthe periodswiththelowesttemperaturesorspringiceduringthe springthaw.Historically,icejamshavebeenrecordedonthe BerounkaRiver(e.g.,in1940,1942and1947;Matous ek, 2004)thatlocallyincreasedthewaterlevelbymorethan5m. Duringthesummerfloods,theriver-waterdensityisusually lowerthanthatofthecave-lakewater,andtheriver floodwatersmostlypenetrateintothecaveintheupperpart ofthewatercolumnofthefloodedsection.Thewaterof winterfloodscanbemoredensethanthewaterofcavelakes andthuspenetratespreferentiallyintotheirdeepestparts. Thebehavioroffloodwatersderivedfromsummerfloods andwinterfloodsrelatedtoicejamsmightthereforebe differentinthefloodedcavesections. Duringtheglacialperiods,increasesinthewaterlevelin therivercausedbyicejamscouldhavebeenmuchmore frequentthaninthecontemporaryclimate.Thedifference inthewaterlevelabovetheicejamandbelowitcanbe severalmeters.Thiscaninducestronginjectionsofriver watersintothekarstandeventheformationofkarstflow pathsinthelimestonemassifparalleltotheriver. C ONCLUSIONS Basedontheresultspresentedinthisstudy,thefollowing conclusionscanbereachedregardingtherelationships Figure13.Conceptualmodelsofflowbetweentheriverandlimestoneaquiferduring,top,normalriverstageand,bottom,a highfloodthatreachesentrancesabovethealluvium.Thecaveontheleftispredominantlyepiphreaticandhasentrances directlyaccessibletofloods;thecaveontherightispredominantlyphreaticandconnectstotheriveronlythroughalluvium. H.V YSOKA ,J.B RUTHANS ,K.Z A K ,J.M LS JournalofCaveandKarstStudies, April2012 N 79

PAGE 82

betweenthecavelakesandtheadjacentmedium-gradient BerounkaRiver: Duringminorwaterlevelpeaksintheriverofupto 1.5mabovenormal,awaterleveloscillationinthekarst aquiferisinduced.Theoscillationamplitudeisslightly dampenedanddelayedcomparedtothatintheriver.Based onthegroundwaterchemistry,itcanbeconcludedthatthe riverwaterdidnotflowmorethanafewtensofmetersinto thekarstaquiferbehindtherockyslopesoftheriver canyonanddidnotreachthesampledcavelakes.Low bicarbonatecontentsandhigh d 13 Cvaluesindicatethat someofthecavelakes’watersarestagnantandundergo CO 2 degassingaccompaniedbycalciteprecipitation. Duringmajorriverfloods(recurrenceinterval 100years),theriverwaterrapidlyfloodsthecaves throughopeningsinthesidesoftherivercanyon,while thoseconnectedtotheriveronlyviathealluviumare floodedbytheelevatedgroundwaterlevel,andtheincrease inthewaterlevelinthemisonlyabout50%oftheincrease intheriverlevel.Asimplehydraulicmodelwassuccessfully usedtosimulateandexplainthewatertableoscillationsin thecavelakes. Thankstotheverylowgradientoftheriver,theriver watercannotinfiltrateintothelimestoneaquiferandflow thereparalleltothecourseoftheriver.Instead,riverwater isinjectedintotheaquiferonlyduringfloodeventsbefore andduringthepeak.Thenthewaterdrainsbackintothe river.ThisprocessissubstantiallyimpededbylowpermeabilitylateHolocenefine-grainedfluvialsediments thatcapthecoarsegravelsintheriverfloodplain. Increasesintheleveloftheriverbyicejamswere recognizedasonefactorthatcanbeimportantin speleogenesis.Undertheseevents,temporaryflowofriver waterthroughthekarstaquiferparalleltotheriveris possiblebecauseoftheveryhighreliefacrosstheicedam. Fastspeleogenesisbymeansoffloodinjectioncouldbe expectedinperiodswhentherivercanyonwasbareor filledbygravelalone.Suchconditionscharacterizedthe lasttwoglacialperiodsandtransitiontotheHolocenein thestudiedrivervalley. A CKNOWLEDGMENTS Theresearchwasfinancedbyresearchprojects MSM00216220855andGAUR80509(CharlesUniversity inPragueandInstitutionalResearchPlan).TheparticipationofKarelZ a kwassupportedbyresearchprogram No.AV0Z30130516,byprojectGACRP210/10/1760,and byprojectSP/2e1/153. R EFERENCES Anderson,R.S.,andAnderson,S.P.,2010,Geomorphology:TheMechanicsandChemistryofLandscapes :CambridgeUniversityPress,640p. Bailly-Comte,V.,Jourde,H.,andPistre,S.,2009,Conceptualizationand classificationofgroundwater-surfacewaterhydrodynamicinteractionsinkarstwatersheds:caseofkarstwatershedoftheCoulazou river(SouthernFrance):JournalofHydrology,v.376,p.456–462. doi:10.1016/j.jhydrol.2009.07.053. Bailly-Comte,V.,Martin,J.B.,Jourde,H.,Screaton,E.J.,Pistre,S.,and Langston,A.,2010,Waterexchangeandpressuretransferbetween conduitsandmatrixandtheirinfluenceonhydrodynamicsoftwo karstaquiferswithsinkingstreams:JournalofHydrology,v.386, p.55–56.doi:10.1016/j.jhydrol.2010.03.005. Bosa k,P.,1998,TheevolutionofkarstandcavesintheKone prusyregion (BohemianKarst,CzechRepublic),PartII:Hydrotermalpaleokarst: ActaCarsologica,v.27,no.2,p.41–61. Bosa k,P.,C lek,V.,andBedna r ova ,J.,1993,Tertiarymorfogenyand karstogenesisoftheBohemiankarst,Prague,TheCzechSpeleological Society,v.21(Karstsediments),p.10–19. Bruthans,J.,andZeman,O.,2000,Newfindingsonhydrogeologyofthe BohemianKarst:C esky kras,v.26,p.41–49.(inCzechwithEnglish abstract). Bruthans,J.,andZeman,O.,2001,Newdataoncharacterandevolution ofundergroundkarstformsintheBohemianKarstandotherareas withdiffuserechargemodeintheCzechRepublic:C esky Kras,v.27, p.21–29.(inCzechwithEnglishabstract). Bruthans,J.,andZeman,O.,2003,Factorscontrollingexokarst morphologyandsedimenttransportthroughcaves:Comparisonof carbonateandsaltkarst:ActaCarsologica,v.32,no.1,p.83–99. Bundschuh,J.,1997,Temporalvariationsofspringwatertemperaturesin relationtotheextentsofheattransportmodesoccurringinthe karstifiedlowerGypsum-Keuperaquifer(Karnian,southernGermany), in Proceedingsofthe12thInternationalCongressof Speleology,ConferenceonLimestoneHydrologyandFissured Media,6th,Switzerland,Volume2,p.129–132. Chlupa c ,I.,Brzobohaty,R.,Kovanda,J.,andStra n k,L.,2002, Geologicka minulostC eske republiky,Prague,Academia,436p. Choquette,P.W.,andPray,L.C.,1970,Geologicnomenclatureand classificationofporosityinsedimentarycarbonates:American AssociationofPetroleumGeologistsBulletin,v.5,no.2,p.207–250. C lek,V.,Dobes ,P.,andZ a k,K.,1994,Formationconditionsofcalcite veinsinthequarry‘‘VKozle(HostimI,Alkazar)’’intheBohemian Karst:JournaloftheCzechGeologicalSociety,v.39,no.4, p.313–318. Doctor,D.H.,Calvin,A.E.,Petric ,M.,Kogovs ek,J.,Urbanc,J.,Lojen, S.,andStichler,W.,2006,Quantificationofkarstaquiferdischarge componentsduringstormeventsthroughend-membermixinganalysis usingnaturalchemistryandstableisotopesastracers:Hydrogeology Journal,v.14,p.1171–1191. Elleder,L.,2004,FloodsincityofBeroun:C esky kras,v.30,p.59–62.(in CzechwithEnglishabstract). Florea,L.J.,andVacher,H.L.,2006,Springflowhydrographs:eogenetic vs.telogenetickarst:GroundWater,v.44,no.3,p.352–361. Havl c ek,V.,1989,Geologicalmap,Prague,CzechGeologicalSurvey, scale1:250,000,sheetKra lu vDvu r,12–413,1sheet. Hora c ek,I.,andLoz ek,V.,1988,PalaeozoologyandtheMid-European Quaternarypast:scopeoftheapproachandselectedresults:Rozpravy C eskoslovenske akademieve d,r adamatematicky chapr rodn chve d, v.98,no.4,p.1–102. Katz,B.G.,Catches,J.S.,Bullen,T.D.,andMichel,R.L.,1998,Changes intheisotopicandchemicalcompositionofgroundwaterresulting fromrechargepulsefromsinkingstream:JournalofHydrology, v.211,p.178–207. Klimchouk,A.B.,Ford,D.C.,Palmer,A.N.,andDreybrodt,W.,2000, Speleogenesis,EvolutionofKarstAquifers,Huntsville,Alabama, NationalSpeleologicalSociety,521p. Koc ,A.,1991,PalaeomagneticinvestigationoftheBerounhighway section:Antropozoikum,U str edn u stavgeologicky ,v.20,p.103–109. Kovanda,J.,1991,ThesignificanceoftheLowerPleistocenesedimentary complexoftheBerounhighway:Antropozoikum,U str edn u stav geologicky ,v.20,p.129–142. Kovanda,J.,andHerzogova ,J.,1986,Druhe chronologicke paradoxonv Kruhove mlomuuSrbska:C esky kras,v.12,p.59–62. Kukla,J.,andLoz ek,V.,1993,Pru zkumr c n chterasvokol Tet naa ota zkaprvn hor c n hoparadoxon:TheCzechSpeleologicalSociety, v.21(Karstsediments),p.30–40. Martin,J.B.,andDean,R.W.,2001,Exchangeofwaterbetweenconduit andmatrixintheFloridanAquifer:ChemicalGeology,v.179, p.145–165.doi:10.1016/S0009-2541(01)00320-5. R ESPONSEOFTHEKARSTPHREATICZONETOFLOODEVENTSINAMAJORRIVER (B OHEMIAN K ARST ,C ZECH R EPUBLIC ) ANDITSIMPLICATIONFOR CAVEGENESIS 80 N JournalofCaveandKarstStudies, April2012

PAGE 83

Matous ek,V.,2004,Ledovy rez imvodn chtoku :Praha,Vy zkumny u stav vodohospoda r sky T.G.Masaryka,Pra ceaStudie,v.199,p.1–203. M O ECR(MinistryoftheEnvironmentoftheCzechRepublic),2003, FinalreportoftheprojectEvaluationofcatastrophicfloodof August2002andofproposaloftheFloodpreventivesystem,Praha, MZ PC R,unpublishedreport,thefinalversionacceptedbythe GovernmentoftheCzechRepubliconJanuary21,2004. (InCzech). Opsahl,S.P.,Chapal,S.E.,Hicks,D.W.,andWheeler,C.K.,2007, Evaluationofground-waterandsurface-waterexchangesusingstreamflowdifferenceanalyses:JournaloftheAmericanWaterResources Association,v.43,no.5,p.1132–1141.doi:10.1111/j.1752-1688.2007. 00093.x. Palmer,A.N.,1991,Originandmorphologyoflimestonecaves: GeologicalSocietyofAmericaBulletin,v.103,p.1–21.doi: 10.1130/0016-7606(1991)103 0001:OAMOLC 2.3.CO;2. Parkhurst,D.L.,andAppelo,C.A.J.,1999,User’sguidetoPHREEQC (Version2)–Acomputerprogramforspeciation,batch-reaction,onedimensionaltransport,andinversegeochemicalcalculations,U.S. GeologySurveyWaterResourcesInvestigationsReport99–4259, 312p. Springer,G.S.,Rowe,H.D.,Hardt,B.,Cocina,F.G.,Edwards,R.L.,and Cheng,H.,2009,Climatedrivenchangesinriverchannelmorphology andbaselevelduringtheHoloceneandLatePleistoceneof southeasternWestVirginia:JournalofCaveandKarstStudies, v.71,no.2,p.121–129. Vacher,H.L.,andMylroie,J.L.,2002,Eogenetickarstfromthe perspectiveofanequivalentporousmedium:Carbonatesand Evaporites,v.17,no.2,p.182–196.doi:10.1007/BF03176484. Vc slova ,B.,1980,Silur-devonBarrandienu-II.fa ze-za ve rec na zpra va, Praha,Stavebn geologie,191p. Za hrubsky ,K.,2003,Moz nostivyuz it izotopu uhl ku 14 Ca 13 Cv hydrogeologiiC eske hokrasu[Ph.D.thesis]:Praha,CharlesUniversity inPrague,136p. Z a k,K.,C lek,V.,Danielisova ,A.,Hlava c ,J.,Kadlec,J.,Kyncl,T., Pokorny ,P.,andSve tl k,I.,2010,Holocenesectionintheexcavation forconstructionoftheHy skovHydropowerPlantanditsbearingto theunderstandingoftheBerounkaRiverfloodplainevolution:C esky kras,v.36,p.42–51.(inCzechwithEnglishabstract). Z a k,K.,andElleder,L.,2007,Historyoffloodsinthekarstcanyonofthe BerounkaRiverinsurroundingsofSrbskovillageduringthelasttwo hundredyears:C esky kras,v.33,p.9–15.(inCzechwithEnglish abstract). Z a k,K.,Hlad kova ,J.,Buzek,F.,Kadlecova ,R.,Loz ek,V.,C lek,V., Kadlec,J.,Z igova ,A.,Bruthans,J.,andS t astny ,M.,2001a,Karstic springandcalcareoustufaaccumulationofHoloceneageinSvaty Jan podSkalou(BohemianKarst),Praha,CzechGeologicalSurvey,Special Papers13,no.1,135p.(inCzechwithextendedEnglishsummary). Z a k,K.,Ta borsky ,Z.,Lachmanova ,M.,andPudilova ,M.,2001b,Heavy mineralassemblagesinallochthonousclasticcavesedimentsofthe BohemianKarst:Apilotstudy:C esky kras,v.27,p.5–14.(inCzech withEnglishabstract). H.V YSOKA ,J.B RUTHANS ,K.Z A K ,J.M LS JournalofCaveandKarstStudies, April2012 N 81

PAGE 84

ANEWSPECIESOFNICOLETIIDAE(INSECTA: ZYGENTOMA)FROMKARTCHNERCAVERNSSTATE PARK,ARIZONA L UIS E SPINASA 1 ,R OBERT B.P APE 2 ,A LANNA H ENNEBERRY 1 AND C HRISTOPHER K INNEAR 1 Abstract: Speleonyctaanachoretes ,n.sp.,isdescribedanddifferentiatedfrom S. ozarkensis ,knownfromcavesintheOzarkPlateau.Thenewspecieswascollectedfrom KartchnerCavernsStateParkinArizona.Morphologyandpreliminaryanalysesusing 16SrRNAcorroboratethat Speleonycta mayberelatedto Texoreddellia ,another nicoletiidgenusfromcavesofTexasandnorthernMexico.Generalinformation regardingitsconservationstatuswithinthecommercialcaveisprovided. I NTRODUCTION WhileinsectsofthefamilyNicoletiidaeareamongthe mostimportantandcommonrepresentativesofcaveadaptedfaunaintheneotropics(EspinasaandGiribet, 2009),theyhavealimitedpresenceincavesofnorthern latitudes.Texashasacomplexofatleastsixspeciesinthe genus Texoreddellia (EspinasaandGiribet,2009),and specimenscollectedfromseveralOzarkcavesinArkansas andOklahomawererecentlydescribedunderanewgenus, Speleonycta (Espinasaetal.,2010). ThecavesystemofKartchnerCavernsisinan isolatedoutcropoflimestonenearthebaseofthe WhetstoneMountainssouthwestofBenson,Arizona (Jagnow,1999).TheKartchnerCavernssystemwas discoveredin1974andwaskeptinnearlypristine condition(TuffsandTenen,1999).In1988thestateof Arizonapurchasedthecaveasastateparkanddeveloped itintooneofthetopshowcavesintheUS,with conservationoftheresourceasitstoppriority.Aspartof thepre-developmentstudies,Welbourn(1999)conducted aninventoryoftheinvertebratecavefaunabetween1989 and1991.Inthatstudy,heobservedthreeindividual nicoletiidsandreportedthemas Nicoletia sp.,butstudied themnofurther. UndertheauspicesofArizonaStateParks,weare conductingare-inventoryoftheinvertebratefaunato analyzeitscurrentstatusandidentifychangesthathave occurredintheinvertebrateecologyofthecaveoverthe twentyyearssincetheoriginalinventory,includingtenyears ofpost-developmentvisitation.Severalnicoletiidspecimens werecollected,theirrelativelyabundancewasassessed,and theyaredescribedhereasanewspeciesusingmorphologic andDNAdata.Genusassignmentisalsocorrected,asthey belongto Speleonycta andnotto Nicoletia. M ATERIALAND M ETHODS Dissectionsofaparatypeweremadewiththeaidofa stereomicroscopeandmountedasfixedpreparationswith Cytoseal60solution(Richard-AllanScientific).Theremainingsampleswereleftinvialswithethanol.Illustrations weremadewiththeaidofacameralucidaattachedtoa microscope.Specimenswillbedepositedinacollectionat theAmericanMuseumofNaturalHistoryinNewYork. ADNAsamplewasextractedusingQiagen’sDNEasy TissueKitbydigestingaleginlysisbuffer.Amplification andsequencingofthe16SrRNAfragmentwasdoneasin EspinasaandGiribet(2009)followingstandardprotocols andprimersusedinthepastfornicoletiids.Chromatogramsobtainedfromtheautomatedsequencerwereread andcontigsmadeusingthesequence-editingsoftware Sequencher3.0.Externalprimerswereexcludedfromthe analyses.Sequenceswerealignedandneighbor-joining analysiswasperformedwithClustalW2. Populationdensitywasroughlyassessedbycounting thenumberofanimalsobservedinfourdaysofobservationsdividedbytheareaofthefloor(includinghabitat underrocks)thatwasaccessiblefordirectexaminationin thesmall,approximately20mlongJackrabbitGallery. R ESULTS MoleculardatawereobtainedfromtheKartchner Cavernsnicoletiidandfrom S.ozarkensis .The16SrRNA fragmentswere496bplong(primersexcluded).Usingthese samesequencesof16SrRNAfragmentsoftwenty-three specieswithinthesubfamilyCubacubaninae,Espinasaand Giribet(2009)observedthatpairsofspecimenswithina populationdifferbyanaverageof1.8nucleotides( 6 2.2 stdev;range0to7;n 5 26),by2.3nucleotides( 6 1.9stdev; range1to6;n 5 9)indifferentpopulationsofthesame species,andby54.7nucleotides( 6 9.5stdev;range45to64; n 5 3)amongsisterspecies.Nucleotidealignmentofthe KartchnerCavernsspecimenwith S.ozarkensis showed considerablesequencedifference(65bp,13.1%;Table1), *CorrespondingAuthor,luis.espinasa@marist.edu 1 SchoolofScience,MaristCollege,3399NorthRoad,Poughkeepsie,NY12601 2 UniversityofArizona,DepartmentofEntomology,Tucson,AZ85721 L.Espinasa,R.B.Pape,A.Henneberry,andC.Kinnear–AnewspeciesofNicoletiidae(Insecta:Zygentoma)fromKartchnerCaverns StatePark,Arazona. JournalofCaveandKarstStudies, v.74,no.1,p.82–89.DOI:10.4311/2011jcks0193 82 N JournalofCaveandKarstStudies, April2012

PAGE 85

supportingthehypothesisthattheybelongtoseparate species.Aneighbor-joiningtreeshowedthenewspeciesand S.ozarkensis clusteredtogether.Bothspeciesof Speleonycta arethenshowntobewithinamonophyleticgroupwith Texoreddellia, attheexclusionofothergeneraofthe Cubacubaninaeandinagreementwithprevioushistone sequencedataof S.ozarkensis (Espinasaetal.,2010). Nevertheless,theseresultsshouldbeconsideredaspreliminary.Establishingthephylogenyofthegroupandassessing whether Speleonycta and Texoreddellia shouldbeplacedina newsubfamilyaretheaimofcurrentresearchusingdata fromadditionalmolecularmarkers. N EW S PECIES : Speleonyctaanachoretes Espinasa,Pape, Henneberry,andKinnear(Figs.1,2A–J,3A–F) Material Holotypemale,body12.7mm,tarsus3 rd leg1.4mm. Paratypes:Males7.5,6.5,and6.2mm;Females15,11,and 10.5mm;Juvenile5mm.KartchnerCaverns,Kartchner CavernsStatePark,CochiseCounty,Arizona.31 u 50 9 16 0 N 110 u 21 9 05 0 W.3/14-18/10,1/9/10and11/27/09. Otherlocality Speleonycta sp.-ArkenstoneCave,ColossalCave MountainPark.PimaCounty,Arizona.32 u 02 9 N 110 u 35 9 W.Male9.4mm;Female9.9mm. Description Maximumbodylengthofsamples15mm.Maximum conservedlengthofantennaeandcaudalappendages 15mmand12mmrespectively.Whencomplete,their lengthisonlyslightlylongerthanbody.Generalcolorlight yellowtowhite.BodyproportionsasinFigure1. Headwithmacrochaetaeandmicrochaetaeasshownin Figure2A,withaboutsevenmacrochaetaeonborderof insertionofantennae.Pedicellusofmaleshorterthanfirst Table1.Alignmentofthe16SrRNAfragmentsequencesofS peleonyctaozarkensis and S.anachoretes n.sp.showed considerablesequencedifference(65bp;13.1%). L.E SPINASA ,R.B.P APE ,A.H ENNEBERRY,AND C.K INNEAR JournalofCaveandKarstStudies, April2012 N 83

PAGE 86

articleandwithclustersofunicellularglands.Fourventral clustersareborderedwithanotveryconspicuousrowof microchaetaeformingaU(Fig.2B)and,onouterborder, ablade-likecuspnotverysclerotizedandwithmore unicellularglandsatitsbase(Fig.2B-C),similarto Speleonyctaozarkensis. Basalarticlesofantennaeoffemale simple. Mouthpartappendagesshort,especiallywhencompared withothercavenicoletiidspecies.Labialpalpasin Figure2D,apicalarticledistinctlylongerthanwideand longerthanthepenultimatearticle.Penultimatearticlewith anotverydistinctbulgecontainingtwomacrochaetae. Labiumandfirstarticleofthelabialpalpwithmacrochaetae.MaxillaasshowninFigure2E.Lastarticleonly slightlylongerthanpenultimateinlargeindividuals,butin smallonesitcanbe1.5 3 .Apexofgaleawithtwoconules, onelongerthanwideandtheotherslightlywiderthanlong (Fig.2F).Twoteethonlacinia.Mandiblechaetotaxyasin Figure2G,withfiveorsixmacrochaetae.Pro-,meso-,and metanotawithseveralmacrochaetaeonpostero-lateral margins,apartfromseveralsetaeofvariedsizes(Fig.2H). Legslong,hindtibiaabout5 3 longerthanwideandslightly shorterthantarsus(Fig.2I).Juvenileswiththinnerlegs.On largestmale(12.7mm),tibiaofsecondlegwithoutthelarge bulgeormodifiedmacrochaetaefoundin S.ozarkensis Femalelegsalsosimple.Clawsshortandwithahairy appearancesimilarto S.ozarkensis AbdominaltergaasinFigure3A,withmultiple macrochaetaeofvariedsizesontheiredgesand1 + 1 distinctmacrochaetaewithinthesurfaceoftheterga. Speleonyctaozarkensis alsohasthesedistinctmacrochaetae,buttheywerenotmentionedorrepresentedinthe originaldescription.AbdominalsternaasinFigures2Jand 3B.UrosternaIentireandII–VIIsubdividedintotwo coxitesandonesternite.UrosternaVIIIandIXofmale entire.UrosternumIIIandIVofadultmaleapparently withoutmodifications.Posteriormarginofurosternum VIIIofmalestraight,withoutemarginationsorprojections inbetweenthestyletsofthissegment(Figs.3Band3G). UrosternumIXofmaleasinFigures3B(Kartchner specimen)and3G(Arkenstonespecimen).Pointof insertionofparamerainurosternumIXslightlybelow levelofbaseofthestyletsofthissegment.Baseofinternal facesofcoxalprocesseswithoneslightlysclerotized macrochaeta(Figs.3Band3G).Penisandparameraas inFigures3Band3G.Parameraverystout,withadistal semi-eversiblevesicleandwithlongspecializedmacrochaetae,somewhatsimilarto S.ozarkensis .Paramera attainabouthalfthelengthofstyletsIX. StyletsIXstoutandwithoutsmallteethonrobust terminalspine.StyletsIXlargerthanothers,without sensorycones,butwithsomesc lerotizedmacrochaetaeof variedsizes(Fig.3G),ventrallywithaboutfivemacrochaetae(Fig.3B).Otherstyletshaveaterminalspine withsmallteethandwithaboutfourmacrochaetae (Fig.3B).UrotergiteXprotruding,shallowlyemarginate inbothsexes,posteriorangleswithseveralmacrochaetae andafewrelativelystrongsetae(Fig.3D).Lengthof Figure1. Speleonyctaanachoretes n.sp.LivespecimenfromKartchnerCaverns. A NEWSPECIESOF N ICOLETIIDAE (I NSECTA :Z YGENTOMA ) FROM K ARTCHNER C AVERNS S TATE P ARK ,A RAZONA 84 N JournalofCaveandKarstStudies, April2012

PAGE 87

Figure2. Speleonyctaanachoretes n.sp.MaleholotypefromKartchnerCaverns.A,headandantennae;B,basalportionof antennae,ventralview;C,pedicellus,dorsalview;D,labium;E,maxilla;F,apexofmaxilla;G,mandible;H,thoracicnota; I,hindleg;J,urosternaIandII. L.E SPINASA ,R.B.P APE ,A.H ENNEBERRY,AND C.K INNEAR JournalofCaveandKarstStudies, April2012 N 85

PAGE 88

Figure3. Speleonyctaanachoretes n.sp.A–B,D–F,maleholotypefromKartchnerCaverns;C,femaleparatypefrom KartchnerCaverns;G,male Speleonycta sp.fromArkenstoneCave.NotethatpartsBandGareshownatdifferentscales;the sizeofspecimensfrombothlocalitiesareactuallysimilar.A,uroterguiteIII;B,genitalareaofmale;C,genitalareaoffemale; D,urotergiteX;E,spinesoncercus;F,cercus;G,parameraandstyletsIX. A NEWSPECIESOF N ICOLETIIDAE (I NSECTA :Z YGENTOMA ) FROM K ARTCHNER C AVERNS S TATE P ARK ,A RAZONA 86 N JournalofCaveandKarstStudies, April2012

PAGE 89

innermacrochaetaeaboutequaltothedistancebetween them. Cerciofadultmalestraight,basallywithtwoorthree annulislightlywiderthanlong,followedbyaverylong annulusandmanysubequalannuli.Theverylongannuli andthefirstofthesubequalannuliwithsmallsensorypegs. Pegsstartassclerotizedchaetaeandprogressivelybecome thicker(Figs.3E-F).Cerciwithlongtrichobothria (Fig.3E).Appendixdorsaliswithoutsensorypegs.Female cercusandappendixdorsalissimple. Subgenitalplateoffemaleroundedtosub-parabolic (Fig.3C).Inthelargestadultfemale(15mm)andin femalesmeasuring10.5and9.9mmlongtheovipositor surpassesapexofstyletsIXbyhalfthelengthofstylets (Fig.3C),butinthe11mmfemaleitsurpassesbyafull lengthofthestylets.Gonapophyseswithabout13annuli. PostembryonicDevelopment Males12.7,9.4,and7.5mmlonghadpedicelluswith unicellularglandsandthedistinctblade-likecusp,parameraattainingabouthalfthelengthofstyletsIX,and smallspinesoncerci.Pedicellusofmales6.5and6.2mm longhadafewunicellularglandsandtheblade-likecusp wasjustbeginningtoform.Theparameraattainonethird thelengthofstyletsIXandnospinesoncerci. Lengthofovipositorinthefourfemalesof15,11,10.5, and9.9mmlongvariedbetweensurpassingthestyletsIX byhalftoonetimesthelengthofthestylets,and gonapophysishadabout13annuli.Inthejuvenilefemale (5mm),theovipositorisjustbarelybeginningtoform. Etymology Thespeciesname anachoretes isGreek,meaning‘‘one thatretiredfromtheworld’’(ahermitorrecluse),in allusiontolivingaratherisolatedexistencewithoutmuch competitionfromotherspecies.Thespeciesinhabits nutrient-poorenvironmentsincaveswherefewother speciescansurvive.Anadjective. Remarks Speleonyctaanachoretes hasthediagnosticcharacters ofthegenus:noscales,urosternumVIIIofmaleflat posteriorly,withoutemarginationsorprojectionsin betweenthestyletsofthissegment,andparamerawith verylongandspecializedchaetaeandadistalsemieversiblevesicle. Speleonyctaanachoretesshares with S.ozarkensis ,the onlyotherdescribedspeciesinthegenus,thedistinctive blade-likecuspwithunicellularglandsonthepedicellusof adultmales,abdominaltergawith1 + 1distinctmacrochaetaewithinthesurfaceoftheterga,andstyletsIXstout witharobustterminalspine.Thetwospeciescanbeeasily differentiatedbytheovipositoroffemales.In S.anachoretes ,femalesupto15mmlonghaveashortovipositorthat surpassesstyletsIXbyhalftoonetimesthelengthofthe styletsandhaveabout13annulionthegonapophyses.In thelargestdescribedfemaleof S.ozarkensis ,despitebeing smaller(12mm),theovipositorislonger(2 3 thelengthof stylets)andgonapophysisismoresubdivided(15to16 annuli).Malesof S.anachoretes alsohaveproportionally smallerparamera;theparameraofmales7.5to12.7mm longattainabouthalfthelengthofstyletsIX,whilein S. ozarkensis ,males11mmlonghavedistinctlylonger paramerathatattaintheapexofthestylets. Speleonyctaanachoretes canfurthermorebedifferentiatedfrom S.ozarkensis byitsproportionallyshorterlast articleofthemaxillarypalpsinlargeindividuals(slightly longerthanpenultimatearticleversus1.5 3 inS.ozarkensis),lessmacrochaetaeonmandibles(5to6versus7),and longerlegs(hindtibiaabout5 3 longerthanwideversus 3.5 3 ).Finally,thetibiaofthesecondleginthemale holotypeof S.ozarkensis isverystout(2 3 longerthan wide)andhasalargebulgewith3distinctlylong, sclerotized,andcurvedmacrochaetae.Thesemodifications areapparentlyabsentinthenewspecies. D ISCUSSION Currently,observationsof S.anachoretes havebeen mostlyconfinedtotheeasternmostpassagesofKartchner Caverns.Welbourn(1999)foundthemneartheentranceof theRedRiverPassage.Wefoundthemprimarilyinthe JackrabbitGalleryandinthevicinityoftheTarantula Room.Thesethreesitesareinthesamegeneralareaand adjointheBigRoom,oneofthetwolargesectionsshown totouristsatKartchnerCaverns.Asinglespecimen observedintheinterioroftheBigRoommayrepresenta vagrantanimal.Twospecimenshavealsobeenfoundinthe GraniteDellssectionofthecave.Thisareaisapproximately275mwestoftheTarantulaRoomandisnot directlyconnectedbyhumanlyaccessiblecavepassage. Bothareasareequallyclose(approximately15m)tothe surfaceofthehillcontainingthecaveandnearareaswhere thereisaninterfaceofepigeanandhypogeanenvironments.Sofar,thenicoletiidsatKartchnerCavernsappear toinhabittheperipheryofthecaveandhavenotbeen founddeepintheinterior.Thisisinsharpcontrastwiththe long-termobservationdataofapopulationof Speleonycta occupyingArkenstoneCave,locatedintheRincon Mountains,36kmnorthwestofKartchnerCaverns.There, theanimalsoccuronlyinthedeeperreachesofthecave. Regardlessoftheirapparenthabitatdifferences,the ArkenstonespecimensdonotappeartobemoretroglomorphicthanthoseatKartchner.Forexample,thelength ofthelegsisthesame(hindtibiaabout5 3 longerthanwide andslightlyshorterthanthetarsus)atbothlocalities. Despitetheirproximity,bothcavesareinisolatedkarst areaswithnopossibleconnectionbetweenthecaves.Due tothespatialseparationofthepopulationsandsinceboth appeartobecaveadapted,onemightexpectthemtobe differentspecies.However,similarsituationsexistwith othercave-inhabitinginvertebratesinsouthernArizona, L.E SPINASA ,R.B.P APE ,A.H ENNEBERRY,AND C.K INNEAR JournalofCaveandKarstStudies, April2012 N 87

PAGE 90

wherespeciesoccurincavesinwidelyseparated,isolated karstareaswithnoundergroundconnection.Anexample is Sitalcinapeacheyi (Opiliones:Laniatores),atroglophilic harvestman,whichisrecordedfromtheBaboquivari, SantaRita,andRinconMountains(UbickandBriggs, 2008).Additionally,ourexaminationoftheonlytwo availablespecimensfromArkenstoneCaverevealedno distinctivemorphologicalcharacteristicthatwouldreadily separatethepopulationsasdistinctspecies.Itremainsto bedeterminedwhethertheArkenstonepopulationisa distinctspeciesornot. Regionaldesertification(dryingandwarming)ofthe northernChihuahuanandSonorandesertsbeganatthe endoftheWisconsinglacialepisodeabout11ka,and reacheditscurrentconditionabout4ka(VanDevender, 1990).Conceivably,asinglewidely-occurringnicoletiid speciesmayhavebeenpresentintheregionpriorto desertification,whentheclimatewaswetterandcooler. Thespecieswouldlikelyhaveoccupiedbothepigean (surface)andhypogean(subsoilandcave)habitats,with geneflowbetweenpopulationsduringthosetimes. Desertificationlikelyresultedinextirpationoftheepigean populations.Thecavepopulationswouldhavebeen isolatedbyinterveninginhospitablehot,dryterrain, preventinggeneflowbetweenanyremainingpopulations. Extantcavepopulationsintheregiontodaymaythus representrelictualpopulationsofthehistoricspecies. ObtainingfreshsamplesfromArkenstoneCaveto sequencetheirDNAcouldhelpresolvewhetherboth populationsrepresentasinglespeciesortwotroglobitic sisterspecies,andifthelatteristhecase,thetimingoftheir separation. Speleonyctaanachoretes iscurrentlythelargesttroglobitedocumentedfromKartchnerCavernsandarguably oneofthemostinteresting.Itisaprimeexampleofan organismdisplayingtraitsofcaveadaptation:blind, albino,withlongappendagesandsensorystructures.Its distinctivemorphologycanmakeitalluringnotonlyto scientists,butalsotothegeneralpublicwhocometovisit KartchnerCaverns.Atthesametime,ithighlightstheir fragilityandtheneedfortheirconservation. Ourobservationsindicatethatthenicoletiidsappearto inhabitacoreareaaroundtheJackrabbitGallerythatis locatedabout15metersfromthedevelopedtouristtrail. Veryroughestimatesindicatethatthissmallgallerycould supportapopulationofaroundonehundrednicoletiids, sincetherewereabouttwoindividualspersquaremeterin thisarea.Oneofthenicoletiidswascaughtlessthanone meterfromalargespotlightthatfullyilluminatedthearea. Thepresenceoflightdoesnotappeartodisturbthe animals. Whatisthesizeofthenicoletiidpopulationat KartchnerCaverns?Thisisanextremelydifficultquestion toanswerwiththedatacurrentlyavailable,andonlyavery vagueestimatecanbeprovidedatthistime.TheJackrabbit Galleryrepresentslessthanonepercentofthetotal mappedlengthofthecave,andthisareaisestimatedto harboruptoahundredindividuals.Furthermore,thereare manyareaswithinthecavethatare,duetotheirsmallsize, inaccessibletoman,butwhichareaccessibletonicoletiids. Theseareasarenotrepresentedbythemappedportionof thecave,andbecausetheyareinaccessibletohumans,their extentcanonlybesurmised.Evenifotherareasofthecave havemuchlowerdensitiesofnicoletiidsthantheJackrabbitGallery,itisprobablysafetoassumethatthetotal populationforthecaveisatleastinthehundreds. ThecommercialdevelopmentofKartchnerCavernsfor tourismdoesnotappeartohaveadverselyaffectedthe nicoletiidpopulation.Theyappeartocopewellwiththe conditionsofilluminationandlimitedhumanuseoftheir habitat.InthecenterofJackrabbitGallerythereisan environmentalmonitoringstation,oneofaseriesofsuch stationsplacedthroughoutthecavethatareusedbypark rangerstocheckcavemicroclimaticconditionsonaregular basis.Despiteroutinevisitsbytheparkstafftothisarea overtheyears,thespecieshaspersisted.Additionally,the developmentofthecaveincludedtheboringofa horizontalaccesstunnel(TarantulaTunnel)withinwhat wenowconsidertobethecorehabitatareaofthespecies inthecave.Furthermore,averticalshaft(JackrabbitShaft) wasdrilleddownfromthesurfacetoprovideanescape routeforconstructionpersonnelandforthemovingof materialsintothecavepriortocompletionofthetunnel access.Thisshaftisrightintheheartofthenicoletiid habitat,andovertheyearsofcaveconstructioncaused extensivehumantrafficandmovementofequipmentand suppliesthroughthearea.Nevertheless,thegrossnumbers ofanimalsobservedinourstudyarewithinanorderof magnitudeofthoseobservedinWelbournÂ’sstudytwenty yearsago.Ifanything,wearefindingmoreanimalsnow. Butthecautionaryprinciplehastobeusedwhen consideringconservationissuesforthespeciesinthecave. NeitherWelbournÂ’snorourstudyhasbeendesignedto provideaquantitativelyaccurateassessmentofthehealth ofthepopulation,norcanweconfidentlystatewhetherthe populationhasdecreasedorincreasedsincethecommercial developmentofthecave. A CKNOWLEDGMENTS WethankRobertR.Casavant,ResearchandScience Manager,ArizonaStateParks.Hehasbeeninstrumental indevelopingthearthropodre-inventoryprojectat KartchnerCaverns.Hehasprovidedsupportatevery stageofourfieldwork,includingcrawlingwithustothe narrowrecessesofthecave.Hiskindness,humor,and willingnesstohelpandsupportourefforthavenotonly madethisprojectareality,butajoy.WethankSteve Willsey,theprincipalrangerassignedtoourproject,who hasbeenpresentoneachofourresearchtrips.SteveÂ’s knowledgeofthecaveandkeeninterestintheprojecthas madehimanindispensableasset.Hehasdevelopedasharp A NEWSPECIESOF N ICOLETIIDAE (I NSECTA :Z YGENTOMA ) FROM K ARTCHNER C AVERNS S TATE P ARK ,A RAZONA 88 N JournalofCaveandKarstStudies, April2012

PAGE 91

eyeforsmall,crawlingthings.Wealsothankparkrangers AbeRandolph,MaryKumiega,andJonLauderbaugh, whohaveaccompaniedusonsomeofourtrips,andother parkpersonnelthathavehelpedmaketheprojectasuccess. WethankCarlOlsonoftheUniversityofArizona, DepartmentofEntomologyforreviewingthemanuscript. Travelexpensesforoneofus(LE)andthreeofhisstudents (TerrenceTurner,AlannaHenneberry,andMichael Mormando)weresupportedbyaVPAAGrantfrom MaristCollege.Molecularworkwassupportedbythe SchoolofScienceatMaristCollege,andtheassistanceof studentsoftheSpring2010GeneticsCourse(BIO320). R EFERENCES Espinasa,L.,Furst,S.,Allen,T.,andSlay,M.E.,2010,Anewgenusofthe subfamilyCubacubaninae(Insecta:Zygentoma:Nicoletiidae)from cavesinsouth-centralandsouthwesternUSA:JournalofCaveand KarstStudies,v.72,no.3,p.161–168,doi:10.4311/jcks2009lsc0097. Espinasa,L.,andGiribet,G.,2009,Livinginthedark—species delimitationbasedoncombinedmolecularandmorphological evidenceinthenicoletiidgenus Texoreddellia Wygodzinsky,1973 (Hexapoda:Zygentoma:Nicoletiidae)inTexasandMexico, in Cokendolpher,J.C.,andReddell,J.R.,eds.,StudiesontheCave andendogeanFaunaofNorthAmerica,PartV:Austin,Texas MemorialMuseumSpeleologicalMonograph7,p.87–110. Jagnow,D.H.,1999,GeologyofKartchnerCaverns,Arizona:Journalof CaveandKarstStudies,v.61,no.2,p.49–58. Tufts,R.,andTenen,G.,1999,DiscoveryandhistoryofKartchner Caverns,Arizona:JournalofCaveandKarstStudies,v.61,no.2, p.44–48. Ubick,D.,andBriggs,T.S.,2008,TheharvestmanfamilyPhalangodidae. 6.Revisionofthe Sitalcina Complex(Opiliones:Laniatores): ProceedingsoftheCaliforniaAcademyofSciences,ser.4,v.59, no.1,p.1–108. VanDevender,T.R.,1990,LateQuaternaryvegetationandclimateofthe ChihuahuanDesert,UnitedStatesandMexico, in Betancourt,J.L., VanDevender,T.R.,andMartin,P.S.,eds.,PackratMiddens:The Last40,000yearsofBioticChange:Tucson,UniversityofArizona Press,467p. Welbourn,W.C.,1999,InvertebratecavefaunaofKartchnerCaverns, KartchnerCaverns,Arizona:JournalofCaveandKarstStudies, v.61,no.2,p.93–101. L.E SPINASA ,R.B.P APE ,A.H ENNEBERRY,AND C.K INNEAR JournalofCaveandKarstStudies, April2012 N 89

PAGE 92

REGIONALIZATIONBASEDONWATERCHEMISTRYAND PHYSICOCHEMICALTRAITSINTHERINGOFCENOTES, YUCATAN,MEXICO R OSELA P E REZ -C EBALLOS 1 ,J ULIA P ACHECOA VILA 2 ,J ORGE I.E UA NA VILA 1 AND H E CTOR H ERNA NDEZ -A RANA 3 Abstract: Assessingwaterqualityinaquifershasbecomeincreasinglyimportantas waterdemandandpollutionconcernsrise.IntheYucatanPeninsula,sinkholes,locally knownascenotes,arekarstformationsthatinterceptthewatertable.Cenotesare distributedacrossthepeninsula,butareparticularlydenseandalignedalongasemicircularformationcalledtheRingofCenotes.Thisareaexhibitsparticular hydrogeologicalpropertiesbecauseitconcentrates,channels,anddischargesfresh watertowardthecoasts.Inthisstudy,weidentifyspatialandtemporalvariationsin chemicalandphysicalvariablesattwenty-twocenotestoidentifygroupsthatshare similarcharacteristics.Watersamplesfromeachcenotesweretakenatthreedepths(0.5, 5.5,and10.5m)andduringthreeseasons(dry,rainy,andcold-frontsseason).Field measurementsofpH,temperature,electricalconductivity,anddissolvedoxygenwere taken,andtheconcentrationsofmajorions(K + ,Na + ,Mg 2 + ,Ca 2 + ,HCO { 3 ,SO 2 { 4 ,Cl 2 andNO { 3 )werequantified.Identifyingregionsofthecenotesweredonebyapplying multivariatestatisticaltechniques(PCA,PERMANOVA,CAP).Thechemicalvariables revealedspatialtrendsamongthecenotes.Weidentifiedthreemainregions.Region1is associatedwithsea-waterencroachmentandhighlevelsofsulfatethattravelthrough preferentialgroundwaterflowpathsfromevaporitesinthesouthernYucatanPeninsula; Region2isarechargezone,andRegion3ischaracterizedbyseawaterencroachment andbythehighchemicalandphysicalvariabilityassociatedwithgroundwaterflowfrom theeast. I NTRODUCTION Karstaquifersarecharacterizedbyhavingcompactand solublecarbonaterocksinwhichthedissolutionprocess (i.e.,karstification)formsconduitsandcavernsthrough whichgroundwaterflows(Antigu ¨edadetal.,2007; CustodioandLLamas,1983,p.1495;Ferna ndezetal., 2003;Mooreetal.,2009).Rainfallquicklyfiltersthrough karstsurfacefeaturesandenterstheaquifer,leadingtothe storageofenormousquantitiesofwater.Inmostkarst regions,thesereservoirsarevitalwatersourcesforhuman consumption,agriculture,livestockraising,andindustry, amongotheruses.Karstaquifersthereforeconstitutea constraininginputinthedevelopmentofregionsandeven countries(Pachecoetal.,2004). Karstsystems’highpermeabilityallowssubstancessuch asnutrients,metals,hydrocarbons,andbacteriatorapidly entertheaquifer.Thesesubstances,includingcontaminants,aredistributedthroughoutthesubterraneanflow network,fromwhichtheyarefrequentlydischargedinto thesea.Karstsystemswithadirectmarineconnectionare alsoexposedtoaseriousriskofseawaterencroachmentas freshwaterisextractedandsalinewatersteadilyadvances intothesystem(Ferna ndezetal.,2003). Groundwatercontainsdissolvedsubstances,largelyin anionicstate.Majorionsincludesodium(Na + ),calcium (Ca 2 + ),magnesium(Mg 2 + ),potassium(K + ),chloride(Cl 2 ), sulfate(SO 2 { 4 ),bicarbonate(HCO { 3 ),andnitrate(NO { 3 ). Theconcentrationsoftheseionscanbeusedtounderstand thechemistryofgroundwateranditsinteractionwiththe subterraneanenvironment,helpingtoidentifythepossible existenceofchemicalprocesses(CustodioandLlamas, 1983,p.1020). TheYucatanPeninsulaisalargekarstzone,withthemain karstcharacteristicbeingsinkholes.Sinkholesarelocally knownascenotesandcanbefoundacrossthepeninsula,but areparticularlyabundantinasemi-circularformationinthe north-centralportionofYucatanstate.ThisRingofCenotes (RC)istheedgeoftheterrestrialportionofthesurface expressionoftheChicxulubmeteoriteimpactcrater(Penfield andCamargo,1981;Perryetal.,1989;Hildebrandetal., 1995;Perryetal.,1995;Mar netal.,2004).TheRCwas proposedin2008asaRamsarsiteforconservationand rationaluse(GobiernodelEstadodeYucatan,2008). *CorrespondingAuthor,rosela.perezc@gmail.com 1 CentrodeInvestigacio nydeEstudiosAvanzadosdelInstitutoPolite cnico NacionalUnidadMe rida,AntiguaCarretaaProgresoKm.6,AportadoPostal 73,Cordemex,97310,Me rida,Yucata n,Me xico 2 UniversidadAuto nomadeYucata nFacultaddeIngenier a,Av.Industriasno ContaminantesporPerife ricoNorteApartadoPostal150Cordemex,97310,Me rida, Yucata n,Me xico 3 ElColegiodelaFronteraSur,Ave.CentenarioKm5.5.Chetumal,QuintanaRoo Me xico R.Pe rez-Ceballos,J.Pacheco-A vila,J.I.Eua n-A vila,andH.Herna ndez-Arana–Regionalizationbasedonwaterchemistryand physicochemicaltraitsintheringofcenotes,Yucatan,Mexico. JournalofCaveandKarstStudies, v.74,no.1,p.90–102.DOI:10.4311/ 2011es0222 90 N JournalofCaveandKarstStudies, April2012

PAGE 93

TheRChasbeenthefocusofanumberofhydrogeologicalstudies.Usinghydraulicgradients,Mar nand Perry(1994)identifiedtheRCasazoneofhigh permeability,whichtheylinkedtotheChicxulubcrater. Hildebrandetal.(1995)determinedthesizeofthecraterand suggestedthattheformationoftheRCwascloselylinkedto adepressionatthecraterÂ’sedge.Steinichetal.(1996,1997) definedtheRCasasystemwithspecialhydrogeological properties,suchashighpermeabilityandsubterranean-river behaviorthatconcentrates,transports,anddischargeswater towardthecoastnearCelestu nandDzilamdeBravo.Using hydrochemicalandhydrogeologicaldata,theyalsodeterminedthatthezoneisdividedintotwowatershedsnearits centralportion.Finally,theyidentifiedgroundwaterflowas movingfromsoutheasttonorthwest.Perryetal.(2002) reportedthatmostgroundwateroftheYucatanaquiferisin approximatechemicalequilibriumwithcalciteanddolomite,butstatedthatothermineralsweresubsaturated. Usingstrontiumisotopes,Perryetal.(2009)demonstrated thatthewaterarrivingatthewesternedgeoftheRC originatesnearChichancanablagooninthesouthernpartof thepeninsula,flowsthroughapermeablefaultzone manifestedonthesurfacebytheTiculRidge,andfinally dischargesintotheCelestu ncoastallagoon(Fig.1). Thesestudieshaveadvancedthehydrologicand hydrochemicalunderstandingoftheYucatanPeninsula. However,chemicaldataareneededfromtheRCinorder toincreaseourunderstandingofthespatialandtemporal variationsinwaterchemicalcomposition.TheRCoffersa myriadofresearchopportunities,becauseitislocatedinan areainfluencedbylocalfactorssuchashumanlanduse,as wellasregionalfactorssuchassubterraneanflow.The presentstudyÂ’sobjectivesweretospatiallyandtemporally quantifythephysicalandchemicalcharacteristicsofwater inrepresentativecenotesalongtheRC,toidentify similaritiesamongthecenotesbasedonchemicalcompositioninordertoobtainahydrochemicalregionalizationof thearea,andtoestablishcharacteristicvaluesforthese regions. S TUDY A REA TheRCareacoversthenorthwestportionofthe YucatanPeninsula(88 u 30 9 and90 u 30 9 W,20 u 00 9 and 21 u 30 9 N)(Steinichetal.,1996).ThewidthoftheRCis approximately12km,andextendsfromthecoastallagoon ofCelestu ninthewesttotheBocasdeDzilamlagoonin theeast,bothinYucata nstate(Fig.1). Figure1.StudyareainYucatanstate,Mexico,showinggroundwatermainflowdirections,isopotentiometriclevels(msal) andspatialdistributionofthetwenty-twosampledcenotesinthegroupsdiscussedinthetext(adaptedfromSARH,1989and Perryetal.,2002). R.P E REZ -C EBALLOS ,J.P ACHECO -A VILA ,J.I.E UA N -A VILA,AND H.H ERNA NDEZ -A RANA JournalofCaveandKarstStudies, April2012 N 91

PAGE 94

Thecalcareousplainofthepeninsulaisformedof Quaternary(Holocene-Pleistocene)sedimentsintheform oflimestonerockofthePaleogeneCarrilloPuerto formation.Surfacetopographyisflatwithverysmooth slope(Perryetal.,1995).Solubilityoflimestoneis producedbycarbondioxideoriginatedinthesoil-plant systeminfiltratedbyrainfall(Gaona-Vizcaynoetal.,1980). Secondaryporosity,fractures,undergroundchannels,and cavernsarethemainsourcesoftheaquiferpermeability (Steinichetal.,1996). ThepeninsulaÂ’saquiferisunconfined,withtheexceptionofaslighthydrogeologicalconfinementcreatedbya thin(0.5to1.4m)calichelayeralongthecoastalmargin (Perryetal.,1989).Thestaticlevelvariesfrom1.0meters abovesealevel(masl)atthecoasttoupto10.0masltothe southoftheRC(INEGI,2002).Thepotentiometriclevel variesfromlessthan1.0maslnearthecoastto5.0maslin thesoutheastpartoftheRC(SARH,1989). Habitatbiodiversityinthestudyregionincludesmarine, coastal,andinlandareas.Marineareascontainfeatures suchasthecontinentalplatform,extensiveseagrassmeadows,andintertidalzones.Thecoastalareaincludes mangrovewetlands,coastaldunes,andseasonallyflooded forest.Inlandareasarecharacterizedbythornydryforest, semi-evergreentropicalforest,anddrytropicalforest.Soil typesintheareaincluderegosolsassociatedwithsand barrierislandsandbeaches,solonchaksandhistosolsinthe mangrovewetlands,andlitosolsandrendzinsintheinland forests(Batllori-Sampedroetal.,2006). ClimatezonesintheYucatanincludethedriestofthe aridandsemiaridclasses,andcoveraspectrumfromthe driestofthesubhumidhottoveryhotclasses.Three seasonsoccurintheregion,adryseasonwithhigh temperaturesandlowrainfall(MarchtoMay),arainy seasonwithfrequentrainfall(JunetoOctober),andacoldfrontsseasonwithwinterstormsandoccasionalrainfall (NovembertoFebruary)(Schmitter-Sotoetal.,2002). Thecenotesincludedinthestudywereopenair,withthe exceptionoftwosemi-openones(seeclassificationproposed byDuch,1991).Cenotesintheareaareusedtoprovide waterforlivestock,cropirrigation,andhumanuse. M ETHODS BasedonthecenoteinventoryoftheYucatanState SecretariatofEcology(SECOL,1999),apreliminary selectionoffifty-twocenoteswithintheRCwasmade.Field dataonaccessibilityanddepthwerecollected.Froma selectionofaboutthirtycenotesreasonablywelldistributed inthestudyarea,twenty-ninewerefoundtohaveaminimum waterdepthof11m.Sevenprovedtobeinaccessible,giving afinalsampleoftwenty-two(Fig.1).Thenamesofthe cenotes,theirmunicipalities,geographiccoordinates,and theirdistancesfromtheshorelinearegiveninTable1. Table1.Namesandlocationsofthetwenty-twosampledcenotes.Thedistancefromshorelineisalongthepathofflow,notthe shortestline. IDNameMunicipalityLatitude,N( u )Longitude,W( u )DistancetoShoreline,km 1Sabtu nCelestu n20.85026090.23559013.45 2Xelactu nKinchil20.88964090.08106026.94 3ChunchucmilCelestu n20.81305690.19666719.29 4ChenhaKopoma 20.68948089.87589059.72 5YaxhaAbala 20.67264089.77415768.13 6KankirixcheAbala 20.63723089.63298082.22 7SabakjaSacalum20.58049089.58820093.10 8NayahTecoh20.64651089.404670108.12 9X-pakayTekit20.53915089.365040115.82 10UitzanTekit20.58069089.342140116.67 11Chonquila Tecoh20.63776089.344290118.23 12Lumha Tekit20.60223089.260180121.69 13UaymilHomu n20.68191089.253120106.74 14IxinhaHuh 20.63035089.110750110.29 15X-colacIzamal20.90977088.86629061.91 16Hotzo Izamal20.90407088.86173062.36 17ChenVa zquezBuctzotz21.14838088.65785041.56 18SanPedroBuctzotz21.18907088.65962036.16 19ItzincabBuctzotz21.22751088.67929025.95 20Dzonotsa bilaBuctzotz21.31051088.74779012.98 21X-kayBuctzotz21.30679088.78810013.14 22DzonotTrejoDzilamGonza lez21.31206088.66786016.41 R EGIONALIZATIONBASEDONWATERCHEMISTRYANDPHYSICOCHEMICALTRAITSINTHERINGOFCENOTES ,Y UCATAN ,M EXICO 92 N JournalofCaveandKarstStudies, April2012

PAGE 95

Atotalof198watersampleswerecollectedfromthe twenty-twocenotes.Watersamplesweretakenatthree waterdepthsduringthethreeseasons,dry,rainy,andcoldfronts.Insitumeasurementsofwatertemperature,pH, electricalconductivityanddissolvedoxygenwereobtained usingamultiparameterprobe(HydrolabG-Quanta). Watersamplesweretakenatdepthsof0.5,5.5,and 10.5musing1.0Land0.25Lplasticbottlesandwerekept coldduringtransporttothelaboratoryformajorionand otherchemicalanalyses.Concentrationsofmajorionsand alkalinityweredeterminedinaccordancewithstandard methodsforeachanalyte(APHA-AWWA-WPCF,2005) (Table2).Qualityoftheresultswasestablishedbyanalysis atleastthreetimesforeachsamplefortitration,titrimetric, andargentometricmethods,andbycalibrationwith standardsandanalysisofreagentblanksforturbidimetric, atomic-absorptionspectrometric,andultravioletspectrophotometricscreeningmethods.Charge-balanceerrors wereastotalcationsminustotalanionsdividedbytotal ions,allinmeqL 2 1 ,times100(Deutsch,1997).Atotalof 87%ofthesampleshadcation/anionbalanceof 6 5%, whiletheremaining13%werebetween 6 5and 6 10%. Tovisualizeandexplorethespatialandtemporal behaviorofthechemicalandphysicalcharacteristics,aset ofplotswerecreated(Figs.2–4).Intheseplotsthe numbersontheabscissaidentifythecenotesaccordingto theirlocationwithintheRCfromCelestu ntoDzilamde Bravo(Fig.1).Thetrendlinesweregeneratedusinga locallyweightedregressionmethod(LOWESS)thatfilters thedispersionofeachvariable(DiRienzoetal.,2008). Anexploratoryanalysisofnormalitywasdonetodecide whichstatisticaltechniquetoapply.Thedatadidnotfita normaldistribution,andtheywereconsequentlyanalyzed withanon-parametricKruskal-Wallistesttoidentifyany significantdifferencesamongdepthsandseasons.Consideringtheunivariateexploratorygraphicanalysis,thelimited dataprovidedbythenon-parametricunivariatestatistical analysis,andthedepthandseasondata,itwasdecidedto regionalizethecenotesusingspatialstructuralcharacteristicstoprovideamoreinformativeandrelevantresult.Three groupswereproposedbasedonthemajorionconcentrationsandthevariabilityoftheseparametersoverseasonand space.Thesegroupswereevaluatedandtestedwitha sequenceofmultivariateanalyses,PrincipalComponent Analysis(PCA),PermutationalMultivariateAnalysisof Variance(PERMANOVA),andCanonicalAnalysisof PrincipalCoordinates(CAP),basedontenhydrochemical variables(conductivity,pH,K + ,Na + ,Mg 2 + ,Ca 2 + ,HCO { 3 SO 2 { 4 ,Cl 2 ,NO { 3 )anddistancefromeachcenotetothe shorelinealongthegroundwaterflow. Figure2.Spatialandtemporalpatternsfromfieldvariablesaveragedforthreedepthsandmeasuredatthetwenty-two cenotes.ECiselectricalconductivityandDOisdissolvedoxygen. Table2.Chemicalmethodsusedinthestudy. AnalysisAnalyticalMethods AlkalinityTitration CalciumEDTAtitrimetric MagnesiumCalculation(differencebetweenhardness andcalciumascalciumcarbonate) ChlorideArgentometric SulfateTurbidimetric SodiumAtomicabsorptionspectrometric PotassiumAtomicabsorptionspectrometric NitrateUltravioletspectrophotometricscreening R.P E REZ -C EBALLOS ,J.P ACHECO -A VILA ,J.I.E UA N -A VILA,AND H.H ERNA NDEZ -A RANA JournalofCaveandKarstStudies, April2012 N 93

PAGE 96

PCAwasappliedtoidentifystructuralpatternsinthe multivariatedatamatrixofelevenvariablesmeasuredat thetwenty-twocenotesgroupedintothreeregions(west, center,andeast)inthreeseasons,treatingthemeasurementsatdifferentdepthsasrepetitions.ThePCAwas basedonacorrelationmatrixoflogarithm-transformed (base10)datanormalizedtoeliminatetheinfluenceof differencesbetweenmeasurementunits.Thistechnique aimstoreducedatadimensionalityandtosearchforlinear relationshipsamongtheoriginalvariablesthroughthe creationofnew,independentvariablesthatexplainthe maximumpossiblevariationintheoriginaldata. Thegeneralhypothesisofnodifferencesbetween seasonsandcenotegroupswastestedusingPERMANOVA.Thissimultaneouslyfalsifiestheresponseofthe elevenmeasuredvariablesversusthefactorsseasonand Figure3.Spatialandtemporalpatternsincationconcentrationsaveragedforthreedepthsandmeasuredatthetwentytwocenotes. Figure4.Spatialandtemporalpatternsofanionconcentrationsaveragedforthreedepthsandmeasuredatthetwentytwocenotes. R EGIONALIZATIONBASEDONWATERCHEMISTRYANDPHYSICOCHEMICALTRAITSINTHERINGOFCENOTES ,Y UCATAN ,M EXICO 94 N JournalofCaveandKarstStudies, April2012

PAGE 97

Table3.Chemicalandphysicalparametersofthewaterinthetwenty-twocenotesfromtheRingofCenotes,averagedoverseasonsanddepth,with standarddeviation. IDpH EC ( m Scm 2 1 ) Temperature ( u C) DO (meqL 2 1 ) Na + (meqL 2 1 ) K + (meqL 2 1 ) Ca 2 + (meqL 2 1 ) Mg 2 + (meqL 2 1 ) Cl 2 (meqL 2 1 ) HCO { 3 (meqL 2 1 ) NO { 3 (meqL 2 1 ) SO 2 { 4 (meqL 2 1 ) 16.962775.5627.604.4515.970.307.046.4617.666.200.085.66 0.25543.561.413.664.320.051.271.553.480.730.071.17 26.872867.7826.832.4816.270.386.635.9318.296.560.163.33 0.21646.810.300.325.850.171.391.224.961.080.171.86 37.362602.2228.022.6313.530.267.185.6215.326.010.045.71 0.41177.332.072.053.470.031.050.380.610.780.020.77 48.192588.8828.855.9712.520.246.817.1413.936.090.126.97 1.3655.782.292.792.580.010.710.510.961.130.071.78 56.871931.1127.362.926.630.155.816.039.076.670.144.02 0.3124.210.040.410.380.010.570.720.540.670.050.87 66.601802.2227.343.335.820.146.315.797.946.830.174.56 0.3412.020.050.370.340.0030.650.820.400.230.070.75 76.801737.5628.154.745.570.136.345.697.476.800.144.74 0.37486.721.022.692.500.051.520.873.080.550.062.72 86.701437.1127.003.694.600.114.904.606.597.420.190.70 0.2837.780.141.640.100.010.330.190.160.300.080.09 96.491140.7827.184.272.650.075.024.113.906.940.171.35 0.4021.670.060.810.080.0010.290.050.290.220.080.14 106.651059.2227.172.972.200.074.684.073.357.040.140.70 0.1326.090.091.080.170.020.280.200.110.150.070.11 116.98779.5625.313.540.980.044.173.451.517.040.190.17 0.1615.330.430.520.070.020.310.190.120.430.080.04 126.861030.7826.285.112.110.064.633.953.406.950.140.34 0.3012.600.481.090.230.020.230.250.350.240.050.07 136.87981.5624.483.261.780.314.743.472.757.020.220.39 0.2636.300.780.930.170.010.550.780.050.350.050.08 146.83987.4426.483.372.100.074.623.383.336.350.140.34 0.2425.790.971.690.200.010.220.490.110.300.070.10 157.221239.0028.656.033.720.094.405.325.427.300.110.53 0.38104.432.343.270.480.010.971.630.731.540.070.06 168.831194.4425.393.454.050.133.135.685.786.300.060.41 1.3572.923.733.760.220.011.500.890.410.980.040.12 177.111489.5627.214.544.640.135.104.156.536.780.090.74 0.30199.991.823.020.500.031.230.780.891.210.060.20 187.311060.1127.125.123.220.073.493.964.745.170.050.49 0.4194.281.932.550.520.010.520.440.440.780.060.11 197.381136.2227.282.753.120.114.493.364.705.930.050.46 0.38110.181.202.470.180.010.580.240.480.860.030.16 R.P E REZ -C EBALLOS ,J.P ACHECO -A VILA ,J.I.E UA N -A VILA,AND H.H ERNA NDEZ -A RANA JournalofCaveandKarstStudies, April2012 N 95

PAGE 98

cenotegroupinamultivariateanalysisofvarianceusinga two-factor(seasonandgroup)anda-posterioristructured designovertheEuclideandistances,applyingpermutationalmethods. Finally,aCAPwasdonetomaximizeseparation betweengroups.ThefirstobjectiveoftheCAPwasto identifyaxeswhosedirectionwasfundamentallydifferent fromthemaximumvariationidentifiedwiththePCAand thatmaximizetheseparationofthegroupsproposedand falsifiedwiththePERMANOVA.Thesecondobjective wastodiagnoseandcross-validatetheaprioriassignment ofthecenotestothethreegroupstodeterminewhich groupsweremoredifferentthanothers.Thisanalytical sequencehelpedidentifysimilaritybetweenthechemical characteristicsofthedifferentcenotesandthenpropose chemicalvaluesrepresentativeofeachregion. R ESULTS ChemicalandphysicochemicalwaterparametersobtainedforeachcenoteareshowninTable3.Theupper valueineachcellisthemeanandthelowervalueisthe standarddeviation. F IELD V ARIABLES AnnualaveragepHwas7.15( 6 0.78),withaminimum of5.81(cenote9)andamaximumof11.06(cenote16), bothrecordedintherainyseason.HighpHvaluesare likelyaresultoforganicmatterderivedfromcattle farming,washedbyrainintothecenotes.TheKruskalWallistestshowedsignificantdifferences( p 0.0001) betweenthemediansforseasonsbutnotfordepths.Values forpHincreasedfromtheRCcenter(cenote9)towardsits twoextremes,withthehighestincreasetowardstheeast.In thecenter,pHvalueswereslightlyacid(5.81–6.9)(Fig.2a). Annualaverageelectricalconductivitywas1560.52 m Scm 2 1 ( 6 659.64),withaminimumof763 m Scm 2 1 (recordedincenote11inthecenter)andamaximumof 3250 m Scm 2 1 (recordedincenote1inthewestnearthe coast),bothduringtherainyseason.Overall,ECvalues increasedfromthecenteroutwardsduringallthreeseasons astheresultofionincorporationfromwater-rock interactionandsea-waterencroachment.However,the increaseinECtowardsthewest(2775 m Scm 2 1 atcenote1) wasnotablyhigherthantowardstheeast(1467 m Scm 2 1 at cenote22).Bothcenotesarelocatedapproximately18km fromthecoast.Duringthecoldfrontsseason,EC increasedatbothendsofthering,incenotes1through4 inthewestandincenotes20to22intheeast(Fig.2b).In eachofthetwoarmsoftheRC(consideredseparately), distancefromthecoastwascorrelatedwithEC.Terrain slopeissteeperintheeastandgentleonthewest,resulting inmoresalineintrusioninthewestcomparedtothe east. Averageannualwatertemperaturewas27.02 u C( 6 1.74)withaminimumof22.2 u C(cenote16at10.5m IDpH EC ( m Scm 2 1 ) Temperature ( u C) DO (meqL 2 1 ) Na + (meqL 2 1 ) K + (meqL 2 1 ) Ca 2 + (meqL 2 1 ) Mg 2 + (meqL 2 1 ) Cl 2 (meqL 2 1 ) HCO { 3 (meqL 2 1 ) NO { 3 (meqL 2 1 ) SO 2 { 4 (meqL 2 1 ) 207.081185.2226.480.423.800.224.513.645.396.210.090.67 0.38353.210.370.211.610.040.820.902.020.940.070.22 218.031837.7826.612.847.370.274.165.789.996.420.022.84 1.1179.971.934.001.070.081.160.861.491.310.022.25 227.321467.3327.624.315.090.295.114.097.196.510.080.70 0.9389.981.644.330.660.271.160.820.471.030.050.16 Table3.Continued. R EGIONALIZATIONBASEDONWATERCHEMISTRYANDPHYSICOCHEMICALTRAITSINTHERINGOFCENOTES ,Y UCATAN ,M EXICO 96 N JournalofCaveandKarstStudies, April2012

PAGE 99

depth)andamaximumof32.2 u C(cenote15at0.5m depth),bothrecordedduringthedryseason.Differences werenotsignificantamongseasonsanddepths.Spatially, thecenotesintheRCÂ’scenterexhibitedlesstemperature variabilitythancenotesatitsedges(Fig.2c). ForallcenotesintheRC,annualaveragedissolved oxygen(DO)contentwas3.73mgL 2 1 ( 6 2.57),witha minimumaverageof0.96mgL 2 1 atadepthof10.5m (cenote3)andamaximumof13.16mgL 2 1 at0.5m (cenote22).MedianvaluesforDOweresignificantly differentamongseasons(p 5 0.0325)anddepth(p 0.0001). Ingeneral,DOdecreasedwithdepth.TheDOconcentrationinthewatercolumnwasleastvariableduringthecoldfrontsseason,withthehighestconcentrationslocatedat thecenteroftheRC(Fig.2d). L ABORATORY V ARIABLES Cations AverageannualNa + contentwas5.81meqL 2 1 ( 6 4.84) withaminimumof0.88meqL 2 1 (cenote11)intherainy seasonandamaximumof22.45meqL 2 1 (cenote1)inthe coldfrontsseason.Nodifferenceswereobservedbetween seasonsordepths.Concentrationstendedtoincreasewith depth,mainlyduringthedryandrainyseasons.Cenotes withthemostconstantNa + concentrationwerelocatedin theRCÂ’scenterandeast.Duringallthreeseasons, concentrationsincreasedfromthecentertowardsthe extremes,withthehighestincreaseobservedoverthe approximately10kmdistancebetweencenotes5 (6.6meqL 2 1 )and4(12.5meqL 2 1 )inthewest.Na + concentrationwasdirectlyrelatedtoelectricalconductivity.Generally,thelowestconcentrationswererecorded duringtherainyseason(Fig.3a). AnnualaverageK + levelswere0.16meqL 2 1 ( 6 0.12), withaminimumof0.02meqL 2 1 (cenote11)anda maximumof0.81meqL 2 1 (cenote2),bothrecordedduring therainyseason.Nodifferenceswereobservedbetween seasonsordepths.OverallK + levelsweremostvariable duringtherainyseasonandleastvariableduringthecold frontsseason.Inallthreeseasons,K + concentrations decreasedwithdepth.Concentrationswerelowestinthe RCÂ’scenter(cenote11)andincreasedtowardsthecoasts. Theoneexceptionwascenote13,whereconcentrations surpassed0.3meqL 2 1 inallthreeseasons(Fig.3b). ForCa 2 + ,annualaverageconcentrationwas 5.15meqL 2 1 ( 6 1.40),withaminimumvalueof 1.67meqL 2 1 (cenote16)duringtherainyseasonand amaximumof9.17meqL 2 1 (cenote3)duringthe dryseason.Differenceswerepresentbetweenseasons ( p 0.0001)andbetweendepths( p 0.0354).Thehighest variabilitywasobservedduringthedryseasonandthe lowestduringthecold-frontsseason.Ca 2 + concentration alsoincreasedwithdepth,particularlyduringthedryand rainyseasonsandonbothsidesoftheRC.Concentration variationwithdepthinthewatercolumnwasleastin cenotesinthecenter.Overall,concentrationswerehighest inthecenotestothewestcomparedtothoseinthecenter andeast(Fig.3c). AverageannualMg 2 + concentrationwas4.80meqL 2 1 ( 6 1.40)withaminimumof1.95meqL 2 1 (cenote20)duringthe dryseasonandamaximumof8.90meqL 2 1 (cenote1)during therainyseason.Variabilitywashighestduringthedry seasonandlowestduringthecold-frontsseason.Mostofthe studiedcenotesexhibitedanincreaseinMg 2 + concentrations withdepth,exceptduringthecold-frontsseason.Nodifferenceswereobservedbetweenseasonsordepths.Spatially, concentrationswerehighest(average6.09meqL 2 1 )inthe west(cenotes1to7)comparedtoalltheothercenotes (average4.2meqL 2 1 ).Thelowestvariabilityinconcentrationswasrecordedincenotesinthecenter(Fig.3d). Anions AverageannualCl 2 concentrationwas7.46meqL 2 1 ( 6 4.9)withaminimumof1.37meqL 2 1 (cenote11) duringtherainyseasonandamaximumof21.15meqL 2 1 (cenote2)duringthedryseason.Nodifferenceswere observedbetweenseasonsordepths.Duringthedryand rainyseasons,concentrationsincreasedwithdepth.WatercolumnCl 2 concentrationvariabilitywithdepthwas highestincenotes1,2and7,inthewest.Inallthree seasons,Cl 2 concentrationswerelowestinthecenter (cenote11)andincreasedtothewestandeast.Theincrease wasgreatesttowardsthewest,withanincreasefrom 9.07meqL 2 1 (cenote5)to13.93meqL 2 1 (cenote4)over anapproximately10kmdistance;thisissimilartothe behaviorofelectricalconductivityandsodium(Fig.4a). AverageannualHCO { 3 concentrationwas6.57meqL 2 1 ( 6 0.93)withaminimumof3.76meqL 2 1 (cenote2) duringtherainyseasonandamaximumof9.36meqL 2 1 (cenote15)duringthedryseason.Concentrationswere mostvariableduringtherainyseasonandleastvariablein thecold-frontsseason.Differenceswereobservedamong seasons( p 5 0.022)andamongdepths( p 5 0.0001).The lowestvariabilityinwater-columnHCO { 3 concentrations wasrecordedinthecentralcenotes.Concentrations generallyincreasedwithdepth.Inallthreeseasons, concentrationswerelowertothewestandeastcompared tohigherlevelsinthecenter(Fig.4b). AverageannualSO 2 { 4 concentrationwas2.08meqL 2 1 ( 6 2.36)withaminimumof0.12meqL 2 1 (cenote11) duringthedryseasonandamaximumof9.27meqL 2 1 (cenote7)duringtherainyseason.Nodifferenceswere observedbetweenseasonsordepths.Variabilitywaslowest duringthecoldfrontsseasonandhighestduringtherainy season.Concentrationwaspositivelyrelatedtodepth.The highestwater-columnconcentrationvariabilitywasrecordedincenotesinthewest.Inallthreeseasons,SO 2 { 4 concentrationsincreasedfromthecentertowardsthewest; anotableincreasewasrecordedovertheapproximately 23kmdistancebetweencenotes9(1.35meqL 2 1 )and7 (4.74meqL 2 1 )(Fig.4c). R.P E REZ -C EBALLOS ,J.P ACHECO -A VILA ,J.I.E UA N -A VILA,AND H.H ERNA NDEZ -A RANA JournalofCaveandKarstStudies, April2012 N 97

PAGE 100

AverageannualNO { 3 concentrationwas0.12meqL 2 1 ( 6 0.08),withaminimumof0.009meqL 2 1 (minimum detectablelevel)duringthedryandcold-frontsseasonsand amaximumof0.56meqL 2 1 (cenote2)duringtherainy season.Differenceswereobservedbetweenseasons ( p 0.0001),butnotbetweendepths.Concentrationswere highlyvariableduringallthreeseasonsandinallthe studiedcenotes,althoughthehighestoverallconcentrationswererecordedintherainyseason.Duringallthree seasons,NO { 3 concentrationswerehighestincenotesinthe centercomparedtothosetothewestandeast(Fig.4d). R EGIONALIZATIONOFTHE R INGOF C ENOTES Threegroupsofcenoteswereestablishedusingtheten hydrochemicalvariablesandthedistancetotheshoreline alongthegroundwaterflowdirections.Group1includes cenotesontheRC’swestside,cenotes1to7.Group2 encompassescenotes8to14inthecenter,andGroup3 includescenotes15to22ontheeastside(Fig.1). ThefirsttwoprincipalcomponentsfromthePrincipal ComponentAnalysisexplained69.3%ofthetotalvariation basedonthemaximumvariationfromelevenoriginal variables(Table4).Thefirstcomponent(PC1)explained 50.1%ofthevariation;themaximumvariationinPC1 separatedGroup1fromtheothersasafunctionof increasesinelectricalconductivityandtheNa + ,Cl 2 and SO 2 { 4 ions.Thesecondcomponent(PC2)explained19.2% ofthevariation;themaximumvariationofPC2separated Groups2and3,mainlyinresponsetotheCa 2 + ,and HCO { 3 ionsandthepHanddistance-to-shorelinevariables.Bothprincipalcomponentsshowedhighvariability withinGroup3(Fig.5). ThePermutationalMultivariateAnalysisofVariance quantifieddifferencesbyseasonandgroup,aswellasthe season-groupinteractionfactor(S 3 G).Therewereclear differencesduetothemain-factorseffect,althoughthe interactiontermbetweenmainfactorswasalsosignificant, meaningtheeffectsofthemainfactorswerenot independent(Table5). Theaposterioripairedpseudo-ttestallowedusto identifytheinteractionsource.Whenthe‘‘group’’effect wasfixed,thepatternofdifferencesbetweenseasons changeddependingonthegroupobserved,butwhenthe ‘‘season’’effectwasfixed,thepatternofdifferences betweengroupsdidnotchange(Table6).Therefore,the differencefromtheinteractiontermwascausedbythe dependenceofseasononthegroupobserved( p 0.01).In otherwords,seasonhasaneffectonlyonaparticular groupofcenotes,butthespatialpatternofgroup differencesisconsistenttemporally,basedonthehydrochemicalcharacteristics(Table6). TheCanonicalAnalysisofPrincipalCoordinatesalso showedtheeffectofthegroupingofthecenotesbasedon thehydrochemicaldata(Fig.6).Cross-validationanalysis oftheclassificationofcenotesintogroupscorroborated that96.8%ofthecenotesinGroup1werecorrectly allocated,100%inGroup2,and91.7%inGroup3. Onceregionalized,theaverageofthehydrochemical variablesvalues( 6 standarddeviation)werecalculatedfor thethreeregions:Region1,Region2,Region3(Fig.1, Table7). D ISCUSSION S PATIAL T RENDS ThelargenumberofcenotesandspecialhydrogeologicalpropertiesmaketheRCauniqueregionforthestudyof groundwater.Thestudiedchemicalandphysicalvariables exhibitedpatternsandspecificbehavioralongtheRC. Table4.Percentageofvariationexplained,pluscoefficients, forthelinearcombinationofelevenvariablesmakingupthe firsttwoprincipalcomponents(PC)forthetwenty-two cenotessampledduringthedry,rainy,andcold-frontsseasons. VariablePC1 a PC2 a pH0.066 0.406 Conductivity 0.411 2 0.094 Na + 0.415 0.022 K + 0.3100.186 Ca 2 + 0.258 2 0.419 Mg 2 + 0.324 2 0.169 Cl 2 0.416 0.029 HCO { 3 2 0.025 2 0.518 SO 2 { 4 0.371 2 0.154 NO { 3 2 0.116 2 0.377 Distancetoshoreline 2 0.248 2 0.390 a BoldnumbersarecoefficientsforthemostinfluentialvariablesoneachPrincipal Component. Figure5.MultivariateordinationfromPrincipalComponentsAnalysis(labelsontheprincipalaxesPC1andPC2 correspondtothemostinfluentialvariables). R EGIONALIZATIONBASEDONWATERCHEMISTRYANDPHYSICOCHEMICALTRAITSINTHERINGOFCENOTES ,Y UCATAN ,M EXICO 98 N JournalofCaveandKarstStudies, April2012

PAGE 101

ThevariablesofpHandtemperaturearedistinctly differentinthecenter,wherepHwasmoderatelyacidic (5.81to6.90)andtemperatureremainedalmostconstant (26.27 u C 6 1.08)duringallthreeseasons.Thissuggests thatthisportionoftheRCfunctionsasahydrological rechargezonethatreceiveswatermainlyfromdirect filtrationofrainfallthroughthepermeablekarstformation.Thisagreeswiththefactthataquiferrechargezones canexhibittemperatureeffectsthatfacilitategreater limestonedissolution(CustodioandLlamas,1983, p.1017),andthelowpHvaluesresultfromthedirect rechargebyrainwaterinfiltration,whichhasameanpH valueof6inthegroundwaterofYucata nstate(Cabrera etal.,1996). Electricalconductivity,Na + andCl 2 concentrations hadsimilarspatialpatterns,withconsiderableincreasesin thewest.Allthreeincreasednotablyfromcenote4to cenote1.Averagevaluesatthesethreecenotes(EC 5 2,708 m Scm 2 1 ;Na + 5 14.58meqL 2 1 ;Cl 2 5 16.30meqL 2 1 )werehigherthannormalfreshwatervalues forthesevariables,whichareEC 5 100to2,000 m Scm 2 1 (Donado,1999),Na + 5 4.35to6.52meqL 2 1 (Custodio andLlamas,1976,p.203),andCl 2 5 0.85–4.22meqL 2 1 (MelloulandGoldenberg,1998).Thecauseofthesehigher values,withCl 2 8.5meqL 2 1 upto55kminlandfrom thecoastatcenote4,isprobablyseawaterencroachment (Escoleroetal.,2005). K + concentrationtendedtoincreasefromthecenter towardsthecoast.K + concentrationsintheaquiferare mainlyduetoagriculturaldischargeandseawaterencroachment(Pachecoetal.,2001).Variabilitywashighest duringtherainyseasonasaresultofrainfallwashing contaminantsintothegroundwater.MostofthegroundwaterinYucata nstatecontainsK + levelsbetween1ppm (0.0256meqL 2 1 )and5ppm(0.128meqL 2 1 ),andalmost alwaysbelow10ppm(0.256meqL 2 1 )(Cabrera,1986). Cenote13hadahigherK + concentration(0.3meqL 2 1 )in allseasonsprobablyduetotheuseoffertilizersinthe Table5.PermutationalMultivariateAnalysisofVarianceresultsofsignificantdifferencesinwaterpropertiesforastructured two-factorcrosseddesignofmainfactorseasonandgroupofcenotesandtheinteractiontermseason 3 group. VariationSourcedf a SS b MS c p seudo-F d p (perm) e Season280.43340.2166.6664 0.0001 Group2904.6452.374.975 0.0001 Season 3 Group440.43910.111.6758 0.0298 a df 5 Degreesoffreedom. b SS 5 SumofSquares. c MS 5 SumofMeansquares. d p seudo-F 5 Teststatisticbasedondistancemeasure. e p (perm) 5 Probabilitybasedonpermutations. Boldfiguresarestatisticallysignificantto p 0.05. Table6.Pair-wisecomparisonsfortheinteractionterm season 3 group,whenlevelofgroup(firstcolumn)andseason (thirdcolumn)isfixed(boldfiguresarestatistically significantat p 0.05). Group1 p (perm)SeasonD p (perm) D-R 0.002 1–2 0.001 D-CF 0.009 1–3 0.001 R-CF0.0892–3 0.001 Group2SeasonR D-R 0.010 1–2 0.001 D-CF 0.002 1–3 0.001 R-CF0.0582–3 0.001 Group3SeasonCF D-R 0.001 1–2 0.001 D-CF0.2511–3 0.001 R-CF 0.012 2–3 0.001 D 5 dry,R 5 rainy,CF 5 cold-frontsseason. Numbers1,2and3refertocenotegroup. Boldfiguresarestatisticallysignificantat p 0.05. Figure6.CanonicalAnalysisofPrincipalCoordinatesof thegroupsbasedonhydrochemicalvariables. R.P E REZ -C EBALLOS ,J.P ACHECO -A VILA ,J.I.E UA N -A VILA,AND H.H ERNA NDEZ -A RANA JournalofCaveandKarstStudies, April2012 N 99

PAGE 102

gardenssurroundingit.Likewise,incenote2duringthe rainyseasonsurfaceconcentrationsreached0.81meqL 2 1 probablyasaresultofcontaminantswashedfromafarm adjacenttothesite. Mg 2 + concentrationswerehighestincenotes1through7 inthewest.Thisionoriginatesfromseawaterand dissolutionofdolomiteandevaporites(Custodioand Llamas,1983,p.1012),suggestingthatthehighconcentrationscouldbeduetoseawaterencroachmentorthepresence ofdolomiteandlow-Mgcalcite(Lefticariuetal.,2006). SO 4 2 2 behaviorwassimilartothatofMg 2 + .Sulfate levelsincenote7(4.74meqL 2 1 )were3.5timeshigherthan incenote9(1.35meqL 2 1 ).Thisdisparitycouldbe explainedbythefactthatthewaterinthewestportionof theRCoriginatesinthesouthnearChichancanabLake, whichhashighSO 4 2 2 levels(52.92meqL 2 1 )(Perryetal., 2002).Bothcenotesarelocatedintheareamarkingthe watersheddividewithintheRC.Thishydraulicseparation oftheRCwasidentifiedbySteinichetal.(1996)byusing theSO 4 2 2 /Cl 2 ratio,anaturaltracer,inananalysisof hydrochemicalandhydrogeologicaldatatoidentifyflow directionandthewatersheddivideinthesouthernportion oftheRC.TheSO 2 { 4 levelsobservedinthepresentstudy exhibitedbehaviorsimilartothatofSO 2 { 4 inthatstudy. TheCa 2 + andHCO { 3 ionsformpartofthecarbonate system.Fullyunderstandingtheirbehaviorrequiresspecificanalysesandcalculationofsaturationindicesfor calcite,aragonite,anddolomitetoidentifychemical processes.Spatialpatternsforeachionweredetermined withintheRC.ThehighestCa 2 + concentrationswereinthe west,andHCO 3 2 hadhigherconcentrationsinthecenter. Finally,nitrate-ionconcentrationswerehighestinthe center.ThisportionoftheRCisarechargeareawith dissolvedoxygenlevelsthatfavoroxidationofnitrogenatedspecies,andasaresult,itcarrieslargeamountsof pollutantssuchasNO 3 2 .Overall,numerousdifferences werefoundalongtheRCthatsuggestregionalizationof theaquiferindifferentareas. S EASONAL T RENDS ThevariablespH,temperature,dissolvedoxygen,Ca 2 + HCO { 3 ,andNO { 3 hadsignificantstatisticaldifferences betweenseasons.Furthermore,electricalconductivity, temperature,dissolvedoxygen,Na + ,Cl 2 ,andNO { 3 clearly showeddifferencesinspatialtrendsvaluesbetween seasons.Electricalconductivity,Na + andCl 2 increasedat theRC’sextremes(cenotes1to4inthewest,cenotes20to 22intheeast)duringthecold-frontsseasondueto movementofseawaterintothecoastalaquifercausedby sealevelchanges.Theseawaterintrusionthreatenswater availabilityforhumanconsumptioninthisregion(ValleLevinsonetal.,2011).Arregu n(2008)proposedthatdue tothecharacteristicsofthepeninsula,sea-levelrisecould changetheinterfaceposition,causingitsgradualmigration inland. Duringtherainyanddryseasons,temperaturedecreasedwithdepth,butthisdidnotoccurduringthecoldfrontsseasonbecauseambienttemperatureislow,leading tohomogeneouswatercolumntemperatures.Dissolved oxygenconcentrationswerehigherandhadlessvariability inthecentercomparedtotheedgesduringthecold-fronts seasonbecauselowambienttemperaturefavorsoxygen dissolutioninthewatercolumn.ConcentrationsofNa + andCl 2 decreasedattheRC’sedges(cenotes1to4inthe west,cenotes20to22intheeast)duringtherainyseason duetodilutionfromaquiferrecharge,aphenomenonalso reportedingroundwater(Cabrera,1986;Cabreraetal., 2002).Nitrateconcentrationsincreasedduringtherainy seasonaspollutantswerewashedintotheaquifer(Pacheco andCabrera,1997;Pachecoetal.,2001).Thechemicaland physicochemicalcharacteristicsofthecenotesofthe YucatanPeninsulashowedmarkedtrendsamongseasons. R EGIONALIZATIONOFTHE R INGOF C ENOTES Thedifferencesinchemicalcompositionbetweenwater fromthethreegroupsofcenotesreflectdistinctwater sources.Thefirstprincipalcomponent(PC1)separatedthe Table7.Averagevaluesoverallsamplesfromthecenotesfromeachgroup. VariableGroup1 N Group2 + Group3 ¤ pH7.09 6 0.756.77 6 0.307.53 6 0.92 Conductivity2329 6 5771059 6 1881326 6 287 Na + (meqL 2 1 )10.90 6 5.472.35 6 1.054.38 6 1.50 K + (meqL 2 1 )0.23 6 0.110.10 6 0.090.16 6 0.12 Ca 2 + (meqL 2 1 )6.59 6 1.124.68 6 0.404.30 6 1.19 Mg 2 + (meqL 2 1 )6.09 6 1.033.86 6 0.564.50 6 1.24 Cl 2 (meqL 2 1 )12.81 6 4.973.55 6 1.456.22 6 1.91 HCO { 3 (meqL 2 1 )6.45 6 0.826.96 6 0.416.33 6 1.20 NO { 3 (meqL 2 1 )0.12 6 0.090.17 6 0.070.07 6 0.06 SO 2 { 4 (meqL 2 1 ) 4.99 6 1.870.57 6 0.380.85 6 1.08 N Group1.CenotesattheRC’swestside. + Group2.CenotesatthecenteroftheRC’s. ¤ Group3.CenotesattheRC’seastside. R EGIONALIZATIONBASEDONWATERCHEMISTRYANDPHYSICOCHEMICALTRAITSINTHERINGOFCENOTES ,Y UCATAN ,M EXICO 100 N JournalofCaveandKarstStudies, April2012

PAGE 103

Group1ofcenotesfromtheothertwobasedonincreased conductivityandNa + ,Cl 2 ,andSO 2 { 4 concentrations comparedtoconcentrationsinGroups2and3.High concentrationsofNa + ,Cl 2 ,andSO 2 { 4 atGroup1are associatedwithseawater,indicatingthatthisgroupis characterizedbyseawaterencroachmentthatcanreach 55kminland.Furthermore,highsulfateconcentrationsin thisgroupcanbeexplainedbythepreferentialgroundwaterflowrichinsulfatefromevaporitesinthesouthern YucatanPeninsula(Perry,2002).Positiveandnegative scoresfromtheprincipalcomponenttwo(PC2)separated Group2fromGroup3.Group2consistedofcenotes locatedatlongdistancesfromtheshorelinethatwere characterizedbylowpHvalues,highNO { 3 ,andhighCa 2 + andHCO { 3 concentrations.Theinfluenceofthecarbonate systemisshowedinPC2throughCa 2 + andHCO { 3 ions,as wellaspHvalues(GarrelsandChrist,1965).ThelowpH valuesinthisgroupwerelikelytobearesultofdirect rechargefromrainwater,whosemeanpHvalueinthe regionsisaround6(Cabreraetal.,1996).ThiscorroboratesthatGroup2,inthecenteroftheRC,isinanaquifer rechargezone.ThecenotesinGroup3exhibitedhigh hydrochemicalvariabilityandwereconsequentlydistributedwithinbothprincipalcomponents.Thevariabilityof thisgroupofcenotesislikelycausedbygroundwaterflow fromtheeastnearthecenote17(Fig.1).Cenotesnearer thecoastweremoresimilarinPC1duetoconcentrations ofNa + andCl 2 similartothoseofGroup1.Thedistances totheshorelineshowedaninfluenceonPC2,indicating thattherewasnotadirectrelationshipwithsea-water encroachmentaccordingtothespatialbehaviorofsodium andchlorideionsreinforcingthedifferencesbetweenthe cenotesfromGroups1and3(Figs.2aand3a). ThePERMANOVArevealedthepresenceofan interactionbetweenseasonandcenotegroups.Thepaired testsrevealedadependenceofseasonsongroupsbutnot viceversa,meaninginter-groupdifferenceswerenotseason dependent;andtherefore,regionalizationispresentyear round.Pairedtestsalsodemonstratedthatseawater encroachmentinGroup1andcarbonate-systemcharacteristicsinGroup2areidentifiableinallthreeseasons. ThisisnotthecaseinGroup3,sinceduringthedryand cold-frontsseasonschemicalprocesseschangeinresponse toenvironmentalvariablesandthemorphologicaland structuralcharacteristicsofeachcenote.TheCAPcorroboratedtheexistenceofGroups1,2and3;indeed,theCAP andPCAgraphicresultsseparatedthegroupsinsimilar ways.Thisconfirmsthatgroupseparationwasonlythe productofthemaximumvariationinchemicaland physicochemicalcharacteristics. C ONCLUSIONS Multivariatestatisticaltechniquesweresuccessfully appliedtoregionalizeintothreelargeregionstwenty-two cenotesintheRCintheMexicanstateofYucatanbased onwaterchemistryandphysicochemicalcharacteristics. Region1encompassescenotesintheRC’swesternportion thathaveNa + ,Cl 2 ,andK + concentrationsmainly associatedwithseawaterencroachment,aswellashigh sulfateconcentrationsoriginatinginthesouthernYucatan Peninsula.Region2includescenotesintheRC’scenter withslightlyacidpHvalues,lowerelectricalconductivity, andlowerion(Na + ,K + ,Mg 2 + ,Cl 2 )concentrations, evidenceofweakinfluencefromseawaterandsuggesting thatthisisanaquiferrechargezone.Region3includes cenotesintheRC’seasternportioncharacterizedbyhigher pHvalues,highhydrochemicalvariabilitycausedby groundwaterflowfromtheeastandlessseawater encroachment.TheRCregionsassociatedwithwater characteristicsmayprovehelpfulwhenmakingmanagementandconservationdecisionssuchastherestriction ofhumanactivities,theselectionofwastedisposalareas, andtheprotectionofrechargeareastohelpmovetoa sustainableuseofthewaterresourcesinthisgeographical area. A CKNOWLEDGMENTS ToConsejoNacionaldeCienciayTecnolog a(CONACYT)-CienciasBa sicas,Me xico,forfinancialsupport throughtheprojectCB-2006-01-60126.Theauthorsare gratefultoanonymousreviewerfortheimportantcommentsonthemanuscripts.R.Pe rez-Ceballosacknowledges adoctoralfellowshipfromCONACYT. R EFERENCES Antigu ¨edad,I.,Morales,T.,andUriarte,J.,2007,Losacu feroska rsticos. CasodelpaisVasco:Ensen anzadelasCienciasdelaTierra,v.5, no.3,p.325–332. APHA-AWWA-WPCF,2005,StandardMethodsfortheExaminationof WaterandWastewater:Washington,D.C.,21Edition,1368p. Arregu nCorte s,F.I.,2008,Evaluacio ndelosefectosdelCambio Clima ticoGlobalsobreelCicloHidrolo gicoenMe xico,in Proceedings,XXCongresoNacionaldeHidra ulica:Toluca,Estado deMe xico,Comisio nNacionaldelAgua,p.1–36. Batllori-Sampedro,E.,Gonza lez-Piedra,J.,D az-Sosa,J.,andFeblesPatro n,J.,2006,Caracterizacio nhidrolo gicadelaregio ncostera noroccidentaldelestadodeYucata n,Me xico:Investigaciones Geogra ficas,Bolet ndelInstitutodeGeograf a,UNAM,no.59, p.74–92. Cabrera,A.,1986,Hidrogeoqu micadelaguasubterra neaenacu feros ca rsticos,[M.S.thesis]:UniversidadAuto nomadeYucata n,Me xico, 76p. Cabrera,A.,Pacheco,J.,andCobos,G.,1996,Caracterizacio nqu micade laprecipitacio npluvialdelaciudaddeMe ridaYucata n:Bolet n acade mico,FIUADY,no.30,p.13–19. Cabrera,A.,Pacheco,J.,Cuevas,E.,Ram rez,J.,Comas,M.,and Ca mara,A.,2002,Hidrogeoqu micadelaguaquesubyaceala JAPAYI,enMe rida,Yucata n,Me xico:Ingenier aRevistaAcade mica,v.6,no.3,p.29–40. Custodio,L.,andLlamas,M.,1983,Hidrolog asubterra nea:Barcelona, Spain,Ed.Omega,2359p. Deutsch,W.J.,1997,Groundwatergeochemistry,Fundamentalsand ApplicationstoContamination:BocaRaton,Florida,CRCPress, 221p. DiRienzo,J.A.,Casanoves,F.,Balzarini,M.G.,Gonzalez,L.,Tablada, M.,andRobledo,C.W.,2008.InfoStat,versio n2008,GrupoInfoStat, FCA,UniversidadNacionaldeCo rdoba,Argentina. R.P E REZ -C EBALLOS ,J.P ACHECO -A VILA ,J.I.E UA N -A VILA,AND H.H ERNA NDEZ -A RANA JournalofCaveandKarstStudies, April2012 N 101

PAGE 104

Donado,L.,1999,Hidrogeoqu mica,Hidrogeolog aconAplicacionesy CasosdeEstudiosLatinoamericanos:PublicadoenInternet,Editado porGonzaloPulidoSilva,PhD.Cap tulo3. Duch,G.,1991,Fisiograf adelestadodeYucata n,surelacio nconla agricultura:Edo.Me xico,UniversidadAuto nomadeChapingo, 229p. Escolero,O.,Mar n,L.,Steinich,B.,Pacheco,J.,Molina,A.,and Anzaldo,J.,2005,Geochemistryofthehydrogeologicalreserveof Me rida,Yucata n,Mexico:Geof sicaInternacional,v.44,no.3, p.301–314. Ferna ndez,R.,Baquero,J.C.,Lorca,D.,andVerdejo,J.,2003,Acu feros ka rsticoscosteros,introduccio nasuconocimiento, in Memoriade TIAC’03Tecnolog adelaIntrusio ndeAguadeMarenAcu feros Costeros:Pa sesmediterra neos,InstitutoGeolo gicoyMinerode Espan ayDiputacio nProvincialdeAlicante,p.3–30. Gaona-Vizcayno,S.,Gordillo-deAnda,T.,andVillasuso-Pino,M.,1980, Cenotes,karstcaracter stico:Mecanismosdeformacio n:UNAM, RevistaInstitutodeGeolog a,v.4,no.1,p.32–36. Garrels,R.,andChrist,C.,1965,Solutions,minerals,andequilibria:New York,HarperandRow,450p. GobiernodelEstadodeYucata n,2008,DesarrolloRegionalparael CrecimientoEquilibrado:PrimerInformedeGobiernode2008,92p. Hildebrand,A.,Pilkington,M.,Connors,M.,Ortiz-Aleman,C.,and Chavez,R.,1995,SizeandstructureoftheChicxulubcraterrevealed byhorizontalgravitygradientsandcenotes:Nature,v.376, p.415–417,doi:10.1038/376415a0. InstitutoNacionaldeEstad stica,Geograf aeInforma tica(INEGI),2002, EstudioHidrolo gicodelestadodeYucata n:GobiernodelEstadode Yucata n,Me xico,77p. Lefticariu,M.,Perry,E.C.,Ward,W.,andLefticariu,L.,2006,PostChicxulubdepositionalanddiagenetichistoryofthenorthwestern YucatanPeninsula,Mexico:SedimentaryGeology,v.183,no.1–2, p.51–69,doi:10.1016/j.sedgeo.2005.09.008. Mar n,L.,Pacheco,J.,andMe ndez,R.,2004,Hidrogeolog adela pen nsuladeYucata n, in Jime nez,B.,andMar n,L.,eds.,ElAguaen Me xicoVistadesdelaAcademia:Me xico,D.F.,AcademiaMexicana deCiencias,p.159–176. Mar n,L.,andPerry,E.,1994,Thehydrogeologyandcontamination potentialofnorthwesternYucatan,Mexico:Geof sicaInternacional, v.33,p.619–623. Melloul,A.J.,andGoldenberg,L.C.,1998,Early-indicatorsignalsof groundwatercontamination:thecaseofseawaterencroachment: EnvironmentalGeology,v.33,no.4,p.279–288,doi:10.1007/ s002540050247. Moore,P.,Martin,J.,andScreaton,E.,2009,Geochemicalandstatistical evidenceofrecharge,mixing,andcontrolsonspringdischargeinan eogenetickarstaquifer:JournalofHydrology,v.376,p.443–455. Pacheco,J.,andCabrera,A.,1997,Groundwatercontaminationby nitratesintheYucatanPeninsula,Mexico:HydrogeologyJournal, v.5,no.2,p.47–53,doi:10.1007/s100400050113. Pacheco,J.,Cabrera,A.,andPe rez,R.,2004,Diagno sticodelacalidad delaguasubterra neaenlossistemasmunicipalesdeabastecimientoen elEstadodeYucata n,Me xico:Ingenier aRevistaAcade mica,v.8, no.2,p.165–179. Pacheco,J.,Mar n,L.,Cabrera,S.,Steinich,B.,andEscolero,O.,2001, Nitratetemporalandspatialpatternin12water-supplywells, Yucatan,Mexico:EnvironmentalGeology,v.40,no.6,p.708–715, doi:10.1007/s002540000180. Penfield,G.,andCamargo,Z.,1981,Definitionofamajorigneouszone inthecentralYucatanplatformwithaeromagneticsandgravity: SocietyofExplorationGeophysisists,51 st AnnualInternational meeting,LosAngeles,Abstracts,37p. Perry,E.,Marin,L.,McClain,J.,andVela zquez,G.,1995,Ringof cenotes(sinkholes)northwestyucatan,Mexico:Itshydrogeologic characteristicsandpossibleassociationwiththeChicxulubimpact crater:Geology,v.23,no.1,p.17–20,doi:10.1130/0091-7613(1995) 023 0017:ROCSNY 2.3.CO;2. Perry,E.,Paytan,A.,Pedersen,B.,andVelazquez-Oliman,G.,2009, GroundwatergeochemistryoftheYucatanPeninsula,Mexico: Constraintsonstratigraphyandhydrogeology:JournalofHydrology, v.367,p.27–40. Perry,E.,Swift,J.,Gamboa,J.,Reeve,A.,Sanborn,R.,Marin,L.,and Villasuso,M.,1989,Geologicandenvironmentalaspectsofsurface cementation,northcoast,Yucatan,Mexico:Geology,v.17,no.9, p.818–821,doi:10.1130/0091-7613(1989)017 0818:GAEAOS 2.3. CO;2. Perry,E.,Velazquez-Oliman,G.,andMarin,L.,2002,ThehydrogeochemistryofthekarstaquifersystemofthenorthernYucatan Peninsula,Mexico:InternationalGeologyReview,v.44,p.191–221, doi:10.2747/0020-6814.44.3.191. Secretar adeAgriculturayRecursosHidra ulicos(SARH),1989,Sinopsis Geohidrolo gicadelEstadodeYucata n:Subsecretar adeInfraestructuraHidra ulica,Direccio nGeneraldeAdministracio nyControlde SistemasHidrolo gicos,48p. Schmitter-Soto,J.,Com n,F.,Escobar-Briones,E.,Herrera-Silveira,J., Alcocer,J.,Sua rez-Morales,E.,El as-Gutie rrez,M.,D az-Arce,V., Mar n,L.,andSteinich,B.,2002,Hydrogeochemicalandbiological characteristicsofcenotesintheYucatanPeninsula(SEMexico): Hydrobiologia,v.467,no.1–3,p.215–228,doi:10.1023/A:10149232 17206. Secretar adeEcolog a(SECOL),1999,CenotesyGrutasdeYucata n: GobiernodelEstadodeYucata n:Me rida,Yucata n,Me xico,Ed. CEPSA,159p. Steinich,B.,andMar n,L.,1997,Determinationofflowcharacteristicsin theaquiferofthenorthwesternpeninsulaofYucatan,Mexico: JournalofHydrology,v.191,p.315–331. Steinich,B.,Vela zquez-Olima n,G.,Mar n,L.,andPerry,E.,1996, Determinationofthegroundwaterdivideinthekarstaquiferof Yucatan,Mexico,combininggeochemicalandhydrogeologicaldata: Geof sicaInternacional,v.35,p.153–159. Valle-Levinson,A.,Marin o-Tapia,I.,Enriquez,C.,andWaterhouse,A., 2011,Tidalvariabilityofsalinityandvelocityfieldsrelatedtointense point-sourcesubmarinegroundwaterdischargesintothecoastal ocean:LimnologyandOceanography,v.56,p.1213–1224,doi: 10.4319/lo.2011.56.4.1213. R EGIONALIZATIONBASEDONWATERCHEMISTRYANDPHYSICOCHEMICALTRAITSINTHERINGOFCENOTES ,Y UCATAN ,M EXICO 102 N JournalofCaveandKarstStudies, April2012

PAGE 105

DELINEATINGPROTECTIONAREASFORCAVESUSING CONTAMINATIONVULNERABILITYMAPPING TECHNIQUES:THECASEOFHERRERI ASCAVE, ASTURIAS,SPAIN A.I.M ARI N 1 ,B.A NDREO 1 ,M.J IME NEZ -S A NCHEZ 2 ,M.J.D OMI NGUEZ -C UESTA 2 AND M.M ELE NDEZ -A SENSIO 2,3 Abstract: Diverseapproachesareadoptedforcaveprotection.Oneapproachis delineatingprotectionareaswithregardtotheirvulnerabilitytocontamination.This paperreportsthemainresultsobtainedfromthedelineationofaprotectionzonefor Herrer asCave,declaredofCulturalInterestbytheAsturiasRegionalGovernment, basedonassessingitsvulnerabilitytocontamination.Thecaveissituatedinacomplex karsthydrogeologicenvironmentinwhichgroundwaterflowsfromsouthwestto northeast,followingthebedrockstructure.Astreamflowsinsidethecave,emergingin aspringlocatedtothenortheastofthesystem.Karstrechargeoccursbydirect infiltrationofrainfalloverlimestoneoutcrops,concentratedinfiltrationofsurface runoffinthewatersheddrainingthecave,anddeferredinfiltrationofwaterfrom alluvialbedsdrainedbyinfluentstreams.Thesoilandvegetationcoversarenaturalin themajorityofthetestsite,butlandusesinthewatershed,includingscatteredfarming, stockbreeding,quarrying,andtouristuse,arechangingthenaturalcharacteristicsand increasingthecave’svulnerabilitytocontamination.Theprocedurefollowedfor delineatingprotectionzonesisbasedonthemethodCOP + Kthatisspecifically designedforvulnerabilitymappingofgroundwaterspringsincarbonateaquifers.To coverthehydrologicalbasinincludedinthecave’scatchmentarea,theprotectionzones establishedincludestwodifferentareas,thehydrogeologicalcatchmentbasinand adjacentlandthatcontributesrunoff.Differentdegreesofprotectioninthezoneshave beenproposedtomakehumanactivitycompatiblewithconservationofthecave,and ourresultsshowremarkabledifferencesfromtheprotectionzonepreviouslyproposed forthesamearea. I NTRODUCTION Thereisincreasingpublicconcerntodayregardingthe protectionofournaturalheritage.Cavesrepresentan outstanding,uniqueelementinthenaturallandscapeand aresometimestheeconomicdrivingforceofregions dependentontourism,andtheymayalsomotivatethe creationofprotectedareas,asforexample,S kocjanCaves RegionalPark,Slovenia(OfficialGazetteoftheRepublic ofSlovenia,801-07/94-5/3,1996)andArde `cheGorge NaturalReserve,France(Statutoryordern u 80-27,14 January1980).Therearedifferentwaysofextending officialprotectiontocaves.InEurope,onemethodis declarationasaSiteofCommunityInterestoranAreaof SpecialConservationInterest.Thisispossiblewhenthe caveisnotopentotourismandprotectionisnecessaryto maintainorrestoreanenvironmentforanimalorplant speciesofcommunityinterestthatrequirestrictprotection, suchasbatspecieslistedintheHabitatsDirectiveofthe EuropeanUnion,92/43/CEE.Inothercases,thecave importancemaybelinkedtothepresenceofarchaeological remnants,motivatingitsdeclarationaspartofthenational culturalheritage,orevenasaWorldHeritageSite.For example,AltamiraCavewasawardedthisdistinctionby UNESCO(SC.85/CONF.008/09,1985). Conservation,evaluation,andmanagementarefrequentlyperformedbylocaladministrations,regionalor nationaldepartmentsofculture,orinternationalagencies. Thus,theUnitedNationshaspromoteddifferentinitiatives specificallyfocusedontheprotectionofculturalheritage. However,theseeffortsaremainlyaimedatrestorationand conservationandrarelyfocusondamageprevention (Catanietal.,2002). Theundergroundextentofcavesisusuallynot apparentfromthesurface,andthissometimesleadsto damagebeingdoneunwittingly.Sincecavesformpartof karstaquifers,thepossibilityofsuchdamageisinfluenced bythehydrogeologicalcharacteristicsofkarstenvironmentsthatareespeciallyvulnerabletocontamination (Zwahlen,2004).AnexampleisgivenbySlovenia,a 1 CentreofHydrogeology(CEHIUMA)andDepartmentofGeology,Universityof Ma laga,Spainaimarin@uma.esandreo@uma.es 2 DepartmentofGeology,UniversityofOviedo,Spainmjimenez@geol.uniovi.es, mjdominguez@geol.uniovi.es 3 SpanishGeologicalSurvey,OviedoOffice,Spainm.melendez@igme.es A.I.MarI n,B.Andreo,M.Jime nez-Sa nchez,M.J.DomI nguez-Cuesta,andM.Mele ndez-Asensio–Delineatingprotectionareasfor cavesusingcontaminationvulnerabilitymappingtechniques:thecaseofHerrer asCave,Asturias,Spain. JournalofCaveandKarst Studies, v.74,no.1,p.103–115.DOI:10.4311/2011jcks0197 JournalofCaveandKarstStudies, April2012 N 103

PAGE 106

countrywithalongtraditionofkarstconservation,where thefirstmeasureforcaveprotectiondatesfrom1908 (BadiuraandBrins ek)andinwhichapproximately20%of the7405cavesrecordedinthe2001CaveSurveyhavebeen contaminatedasaconsequenceofhumanactivity(Kepa, 2001). Inrecentyears,spurredbytheworkofinternational agenciessuchasUNESCO,therehasbeenanincreasein legislationandpracticalworkfocusingonpreventionof damagetotheculturalandnaturalheritage.Atthe32nd UNESCOConventionConcerningtheProtectionofthe WorldCulturalandNaturalHeritage,warningsofthe risksfacedregardingtheconservationofcavesweremade: ‘‘Inthecontextofcontemporarydiscoveryandopeningof caves,complexriskshavearisenrelatedtothealterationof thephysical,geologicalandbiologicalconditionsof conservation’’and‘‘Extremelyrigorousmanagementis requiredtoensurethattherisksareeffectivelydealtwith’’ (WHC-08/32.COM/24Rev,2009).Nevertheless,asyetno ruleshavebeenclearlydefinedinlegislationregarding whatkindsofmeasuresshouldbetakenandhowprotected areasshouldbedefinedtoprotectcavesandtheir surroundings.Inmanycases,protectionislimitedto preventingentryintothecave,aninsufficientmeasurethat oftenfails(Lera,2009).Moreover,thecomplexityand diversityofthekarstmediumhampersthedesignofa universalmethodofestablishingprotectedareasforcaves andtheirsurroundings.Accordingly,anyproposalshould considertheidentificationandevaluationofthenatural processesthatinvolveriskstotheconservationofthe cave.Inaddition,itwillbenecessarytoconsiderhuman activitiessuchasconstruction,livestockfarming,and agriculturethatcouldhaveanegativeimpactonthe naturalevolutionofcaves(Sa nchez-Moraletal.,2002). Evaluatingthevulnerabilityofacavetocontamination, onthebasisofthecharacteristicsofthephysicalmediumin whichitiscontained,canbeusedtopromoteland-use managementthatiscompatiblewiththeprotectionand conservationofthecave.Thisapproachisaninitialstepin thedelineationofprotectionzonesforthecaveandin preventingtheriskofcontamination.Furthermore,agood planfortheprotectionandconservationofacavecould includeotherstudies,suchasanalysisofgeologicalrisks, particularlyregardingitsresistancetovibrations(Sa nchez etal.,2007;Iriarteetal.,2010),theevaluationofhuman impactusingquantitativeindexesbasedongeomorphologicalfeatures(Jime nez-Sa nchezetal.,2011),andstudies ofthecave’senvironmentalcapacitywithrespecttothe optimumnumberofvisitspermissibleintouristcaves (Pulido-Boschetal.,1997;Cuevas-Gonza lezetal.,2010). Theconceptofthecontaminationvulnerabilityofan aquiferhasbeendefinedbymanyinvestigators,among themMargat(1968)andZaporoz ec(1994).Theintrinsic vulnerabilityofanaquiferinvolvesitssensitivityto contamination,takingintoaccountitsgeological,hydrologicalandhydrogeologicalcharacteristics,independentof thenatureofthecontaminantsandthecontamination scenario(Zwahlen,2004).Thisconceptofintrinsic vulnerability,proposedintheEuropeanCOSTAction 620,isbasedontheorigin-flow-targetconceptualmodel (Fig.1).Usingthismodel,itispossibletodistinguish betweenresourceprotectionandwatersupplyprotection. Figure1.Diagramoftheorigin-flow-targetconceptualmodel,proposedinEuropeanCOSTAction620(modifiedfrom GoldscheiderinZwahlen,2004). D ELINEATINGPROTECTIONAREASFORCAVESUSINGCONTAMINATIONVULNERABILITYMAPPINGTECHNIQUES:THECASEOF H ERRERI AS C AVE A STURIAS ,S PAIN 104 N JournalofCaveandKarstStudies, April2012

PAGE 107

Consideringresourceprotection,itisassumedthatthe elementtobeprotectedisthewaterstoredintheaquifer belowthepiezometricsurface.Theflowofthepolluting agentfromtheoriginisconsideredtobepractically vertical,passingthroughthesoilandtheunsaturatedzone oftheaquifer.Inthecaseofsourceprotection,thetargetof protectionistheaquiferdischargepointatawellorspring. Conceptually,thecontaminantisassumedtobetransportedfromtheoriginatthesurfacetotheaquiferandfrom theretothedischargepointataspringorwell.This displacementhasaverticalcomponentfromtheoriginto thepiezometricsurfaceandahorizontalone,thelatter beinginthesaturatedzoneoftheaquifer. Themainaimofthepresentstudyistoadaptthe groundwater-contaminationvulnerabilitymappingmethod tothedelineationofprotectionareasforcavesclosetothe piezometriclevelortoaspring.Theexperimentalsite consideredwastheHerrer asCaveinnorthernSpain,for whichaprotectionboundaryhasbeenestablishedbythe CultureandTourismDepartmentoftheAsturiasRegional Government(BOPA,2008;BOE,2010),permittinga comparisonbetweenthisandtheprotectionareaestablishedinpresentwork. T HE S TUDY A REA ThecaveknownasHerrer as,LaHerrer a,orElBola u issituatedinthedistrictofLlanesinAsturias,Spain (Fig.2).Thecavecontainsvaluableexamplesofprehistoricartthatmakeitparticularlyinterestingfroma heritagepointofview.Thecavepaintingsconsistofasetof gridformscomposedofindividualmarksandparallellines, locatedintheroofofthecave.Thesignificanceoftherock art,whichisattributedtotheMagdalenian(Jorda and Mallo,1972),ledtothecaveÂ’sbeingdeclaredaGood ObjectofCulturalInterestasaresultofapplicationofthe 1985SpanishHistoricalHeritageLaw(BOE,1985).The floorplanofthecaveisintricateandirregular,witha horizontalextentoflessthan200mandtwoaccesspoints toamaingallery,thesouthwesternbranchofwhichis blockedbyagatethatprotectstheroomcontainingthe cavepaintings.Landusesurroundingthecaveischaracterizedbyscatteredfarmingandstockbreedingthatcoexist withtouristuse.Also,theremnantsofminingand quarryingworkscanberecognizedinthelandscape, includinghistoricquarryinginthequaternaryalluvial fans,aswellasanoldiron-manganesemine,locatedinthe limestonetothewestofthecave(Dom nguez-Cuestaetal., 2010).Thetopofthekarstmassifinwhichthecaveis locatedispresentlycoveredbyabuildingmaterialdump,a potentialsourceofcavecontamination. Herrer asCaveislocatedinacomplexhydrogeological environmentthatgovernsitsvulnerabilitytocontaminationeventsandthus,itsneedforprotection.Forthis reason,amultidisciplinarystudyinvolvinggeological, geomorphological,andhydrogeologicalaspectshavebeen doneina10km 2 area,includingthecaveÂ’ssurroundings (Jime nez-Sa nchezetal.,2010). Thelandscapesurroundingthecavecontainsmountainousrangesandvalleystrendingeast-west,followingthe geologicalcontactsbetweenbedrockformations.Altitudes descendfromthemountainousrangesinthesouth(Sierra delCuera,1315metersabovesealevel)tothevicinityof theCantabrianshoretothenorth,whereremnantsof karstifiedmarineterracescanberecognizedbetween30 and189masl.Thehydrographicnetworkissuperimposed onthisrelief,withstreamsmainlyflowingfromsouthto north,givingplacetotorrentialbasinsandcoalescing alluvial-fandeposits.HerreriasCaveislocatedatthe bottomofoneofthesetorrentialbasins,whichshowsa totalsurfaceof3.37km 2 ,ameanslopeof23 u (maximumof 57 u ),andaltitudesrangingbetween34mand757m (Fig.2b). Geologically,thestudyareaislocatedinthePongaCueraUnit(Marqu nez,1989).Thestratigraphicsequence iscomposedofBarriosquartziteoftheOrdovicianage,the ErmitasandstoneofLateDevonianage,andseveral Carboniferouslimestoneformations:theAlba,Barcaliente, andCuerathatoutcroptrendingeast-west,showing subverticaldips(Figs.3and4).Themostoutstanding Figure2.(A)LocationofHerrer asCave.(B)Siteofcave andhydrologicalwatershedoverdigitalelevationmodel. A.I.M ARI N ,B.A NDREO ,M.J IME NEZ -S A NCHEZ ,M.J.D OMI NGUEZ -C UESTA,AND M.M ELE NDEZ -A SENSIO JournalofCaveandKarstStudies, April2012 N 105

PAGE 108

structureisasubverticalthrust,alsotrendingeast-west, thatputsOrdovicianquartziteincontactwiththeCuera limestonetothesouth.Thecaveisdevelopedinaregionof Cueralimestonelocatedinthenorthofthestudyarea.The PaleozoicbedrockiscoveredbyseveralQuaternary formationsofalluvial,colluvial,andkarsticorigin. Fourhydrogeologicunitscanbedistinguishedinthe studyarea(Figs.3and4),fromsouthtonorth:Unit1 Figure3.Geological-hydrogeologicalmapofHerrer asCaveanditssurroundings. D ELINEATINGPROTECTIONAREASFORCAVESUSINGCONTAMINATIONVULNERABILITYMAPPINGTECHNIQUES:THECASEOF H ERRERI AS C AVE A STURIAS ,S PAIN 106 N JournalofCaveandKarstStudies, April2012

PAGE 109

comprisestheCueralimestones;Unit2consistsof OrdovicianquartziteandDevoniansandstoneandis separatedofUnit1byathrustfault;Unit3contains CarboniferouslimestoneoftheCueraandBarcaliente Formations.Finally,Unit4ischaracterizedbyQuaternary alluvialmaterialsthatlieunconformablyoverthebedrock formations(Jime nez-Sa nchezetal.,2008a).Units1and3, madeupofkarstifiablelimestone,arecarbonateaquifers; Herrer asCaveislocatedinUnit3,tothenorthofthe thrustfault(Figs.3and4).TheOrdovicianquartziteand Devoniansandstoneconstituteabarrierthatpreventsthe undergroundconnectionbetweenthetwocarbonate aquifers(Ortun oetal.,2004;Fig.4). TheUnits1and2formpartoftherechargeareaofthe aquiferinwhichthecaveislocated(Unit3),becausepart ofthesurfaceareadrainstowardthisaquifervianumerous streams.Unit3,withagroundwaterflownorthward,is drainedbysprings,includingthatofElBola u,whichisthe resurgenceofthiscave.Thegroundwaterflowintothe cavetakesplacethroughmultipledrippointsandviaa subterraneanwatercoursethatisvisibleinsomeareasin thecave.Thus,asshowninFigures2and3,inthestudy areathehydrologicalwatershedandhydrogeologicalbasin donotcoincide. Fromtheaboveobservations,itcanbededucedthat therechargeareaofbothElBola uspringandHerrer as Cave,containstwodifferentsectors.Thesouthernsector constitutesthelimestoneofUnit1andthesandstoneand quartziteofUnit2.Surfacerunoffgeneratedinthissector infiltratesintotheaquiferlimestonesfarthernorthinUnit 3throughtheoverlyingQuaternarydepositsofUnit4. Therefore,theinfluenceofthesouthernsectoronthe cave’svulnerabilitytocontaminationisrelatedtothe infiltrationofsurfacewaterandpotentialcontaminants fromthehydrologicalwatershed.Theconnectionbetween certainpointsofthishydrographicnetworkandthe cavehasbeendemonstratedbytheinjectionofsodium fluoresceineintotheCarroceoRiver(Jime nez-Sa nchez etal.,2010). Thenorthernsectoristhepartoftheaquifercontaining theUnit3limestonethatdrainstowardtheElBola u spring.Despitethelackofdataonpiezometriclevels,it seemsreasonabletobelievethattheElBola uspringdrains waterdirectlyinfiltratingintothelimestonesoutcrops (Unit3),togetherwiththeinfiltrationwaterfromsurface runofffromthesouthernareas.Inadditiontothesetwo components,thereisrechargefromtheQuaternary materialsofUnit4,asalsodemonstratedbytracer injectionintheCarroceoRiver(Jime nez-Sa nchezetal., 2010). Analysesofwatersamplescollectedatadrippointin thecaveduringthemonthsofFebruary–June2007show thattheevolutionoftheNO { 3 ,NO { 2 ,TOC(totalorganic carbon)andCl 2 contentisrelatedtorechargeepisodes, butalsotocontaminationfromhumanactivitiesinthearea ofthecave’srechargezone.Peaksof20mgL 2 1 and Figure4.Geological-hydrogeologicaldiagramofHerrer asCaveanditssurroundings. A.I.M ARI N ,B.A NDREO ,M.J IME NEZ -S A NCHEZ ,M.J.D OMI NGUEZ -C UESTA,AND M.M ELE NDEZ -A SENSIO JournalofCaveandKarstStudies, April2012 N 107

PAGE 110

50mgL 2 1 ofTOC,15.63mgL 2 1 ofNO { 3 and < 17mgL 2 1 ofNO { 2 havebeendetectedinthedripwater(Jime nezSa nchezetal.,2008b). A PPLICATIONOFTHE COP + KM ETHOD Inthisstudy,theCOP + Kmethod(V asetal.,2006; Andreoetal.,2009)wasapplied.Thismethodwas originallydevelopedtoevaluatecontaminationvulnerabilityandtodelineateprotectionareasforwatersupply springsorwellsincarbonate(karst)aquifers.However, thismethodmustbereinterpretedandspecificallyadapted tothiscase,consideringHerrer asCaveitselfisa protectiontarget(Fig.1),inawayanalogoustothat appliedtowatersupplysources. TheCOP + KmethodisanextensionoftheCOPmethod (V asetal.,2006),bywhichthecontaminationvulnerabilityofkarsticaquiferscanbeevaluatedandonthebasisof thisassessment,protectionzonesdefined.Thismethod incorporatestheconceptualmodelandtheindications setoutintheframeworkofEuropeanCOSTAction 620(Zwahlen,2004).TheCOP + Kmethodisbasedon estimatingthesurfaceflowrechargeconditions(Cfactor), theprotectioncapacityofthelayersoverlyingtheaquifer (Ofactor),andtheprecipitation(Pfactor),togetherwith theparticipationofthesaturatedzone(Kfactor).The methodisappliedusingcartographicoverlaycreatedwith geographicinformationsystems.Theprocedure,classification,andscoringofthevariablesusedtoevaluate vulnerabilityaccordingtotheCOP + Kmethodareshown inFig.5andexplainedindetailbyV asetal.(2006)and Andreoetal.(2009). InaccordancewithcurrentknowledgeofthehydrogeologicalsystemofHerrer asCaveanditssurroundings (Jime nez-Sa nchezetal.,2008a,2008b,2010),theprotectionofthecavityshouldconsiderbothsurfacewaterinthe hydrologicalwatershedandgroundwaterinthehydrogeologicalbasin(Figs.3and4).However,applicationof theCOP + Kmethodisonlypossibleinthehydrogeological basinoftheHerrer asCave,wherethereisavulnerable carbonateaquiferanddepositsofalluvialmaterials(Units 3and4,Fig.2).SinceUnit2(Ordovicianquartzitesand Devoniansandstones)doesnotconstituteanaquifer,itis notpossibletoevaluatethethicknessoftheunsaturated zone(Ofactor)orthetransitthroughthesaturatedzoneof possiblecontaminants(Kfactor).Likewise,novaluationis madeoftheKfactor(transitthroughthesaturatedzone) inUnit1because,asremarkedabove,thisunithasno hydrogeologicalconnectionwiththecave.Themainresults ofourevaluationoftheparameterscomposingthemethod aredescribedbelow(Fig.6). CF ACTOR Thisfactorwasthemostdifficultonetoassess,dueto thedoublecomponentofrunoffandinfiltrationactingas rechargeandtothenon-coincidenceofthehydrological andhydrogeologicalwatersheds.Thesurfaceconditions fortheconcentrationofwaterflowstorechargetheaquifer (theCfactor)wereevaluatedbasedonapriorclassification ofthestudyareaintothetwoinfiltrationsourcesdefinedin theoriginalmethod,concentratedinfiltrationviakarstic swallowholesanddirectinfiltrationofrainfallover outcropsintherestofthecatchment(V asetal.,2006). However,anadditionalsourceofinfiltrationaffectsa significantsectorofthestudyarea,deferredinfiltration throughthebedsofinfluentstreamsthatwasdemonstrated bytheuseoftracersatcertainpoints(Jime nez-Sa nchezet al.,2010).EvaluationoftheCfactorofthisnewsourcewas carriedoutusinganadaptationofthetwoscenarios previouslydefinedbyV asetal.(2006).Thescoresofthe sv,dh,anddssubfactorsweremodified(seeTablesXII– XIVinFig.5;scenario3ofCfactor),takingintoaccount thatthisinfiltrationthroughtheunsaturatedzoneis concentrated,butnotasmuchasinsinkingstreamsvia swallowholes. OF ACTOR Thespatialdistributionoftheprotectioncapacityofthe layersoverlyingtheaquiferinwhichHerrer asCaveis developedreflectsthedistributionofthesoilthatis importantinprotectinggroundwateragainstcontaminationepisodes.Abovethisaquifer,thesoilistheonlylayer conferringheterogeneitytotheOfactor,asthelithologyis similarthroughoutthehydrogeologicalbasintobe delineated.Theprotectioncapacityisverylowimmediately surroundingofthecavebecausethesoilispracticallynonexistent.Theprotectioncapacityisslightlyhigherinthe restofthehydrogeologicalbasinwherethesoilisbetter developed,andithasbeenevenconsideredhighinthe smallareaswheresoilofaclayeytextureandgreater thicknesshasbeenmapped. PF ACTOR Thisfactorwasevaluatedusingdataobtainedfromthe ParresdeLlanesmeteorologicalstation,located1.3km northwestofthecave.ThePfactorisbasedonthe quantityandintensityofprecipitationeventsbasedonthe dailyprecipitationofahistoricalsetofwetyears.The averagerainfallis1600mmandtheaveragenumberof daysofrainperyearis135.Thisannualprecipitation helpsinthedilutionofpotentialcontaminantthatislikely tobedominantprocessinsteadofinfiltration(V asetal. 2006).Amoderatereductioninthelevelofprotection providedbytheunsaturatedzone(evaluatedbyOfactor) duetotheprecipitationcharacteristicsinthisareahas beenidentified.Thisparameterwastakenashomogeneousforthetestsite(Fig.6)becauseofthelackofan adequatenumberofmeteorologicalstations.Although thisfactordoesnotinfluencethedistributionofvulnerabilityclasses,itaffectstheindexesofvulnerability, reducingthevalueresultingfromtheCandOfactors;the valueofPfactoris0.7. D ELINEATINGPROTECTIONAREASFORCAVESUSINGCONTAMINATIONVULNERABILITYMAPPINGTECHNIQUES:THECASEOF H ERRERI AS C AVE A STURIAS ,S PAIN 108 N JournalofCaveandKarstStudies, April2012

PAGE 111

TheCOPvulnerabilityindexmap(COPinFig.6),is thecartographicoverlayoftheC,O,andPfactorsand showsthevulnerabilitytocontaminationofthetestsite. TheaquiferinwhichHerrer asCaveislocatedpresentsa contaminationvulnerabilitylevelrangingfromhightovery high.Inonlyafewsmallsectorsisthevulnerability classifiedasmoderate.Thedistributionoftheareaswhere contaminationvulnerabilityisveryhighisdeterminedby Figure5.DiagramoftheCOP + Kmethod,showingthedifferentiationoftheC,O,P,andKfactors.ModifiedfromV as etal.(2006). A.I.M ARI N ,B.A NDREO ,M.J IME NEZ -S A NCHEZ ,M.J.D OMI NGUEZ -C UESTA,AND M.M ELE NDEZ -A SENSIO JournalofCaveandKarstStudies, April2012 N 109

PAGE 112

twofactors:thescantprotectioncapacityaffordedbythe layersoverlyingtheaquifer(verylowOfactorvalues)and theconsiderablereductionoftheprotectioncapacity resultingfromtheconcentrationofsurfacerunoffflows provokinginfiltration(theCfactor).Onlywheretheslope islowerthan8%istheresourcevulnerabilityonthe hydrologicalwatershedhighinsteadofveryhigh.Inthe COPindexmapinFigure6,thewatershedareafedby influentstreamsshowsvaluesofvulnerabilityhigherthan thoseobtainedinsimilarareasoutofthishydrographic basin. KF ACTOR Thisfactorwasevaluatedafterdeterminingthe directionandvelocityofgroundwaterflowinthestudy areausingtracertestsandotherhydrogeologiccriteria. Specifically,threeflowvelocitiesweredetermined(Fig.7). BetweenMeandrosHallandPinturasHallinthecave,the approximatelyeasttowestflowvelocitywas128mh 2 1 whilebetweenPinturasHallandtheElBola uspring,it rangedfrom237to250mh 2 1 inanapproximately northeasterlydirection(twotracertests;Jime nez-Sa nchez etal.,2010).Thetracertestswerecarriedoutinasmall sectorofthehydrogeologicalbasin,andsotheirabilityto representtheentireaquiferremainsuncertainuntil additionaldataareacquiredtoimprovetheKfactor evaluation.Thedifferenceinvelocitiescalculatedusingthe tracertestshasnoimpactontheapplicationoftheCOP + K methodbecausetheentirestudyareahasatransittime(the tsubfactorinthemethod)oflessthanoneday.Ther subfactorconsidersthecontributionandconnectionrates ofdifferentpartsoftheaquifertothecave.Theoverall hydrogeologicalbasinwasconsidereddirectlyconnectedto thecave. Figure6.C,O,P,andCOPmapsassessingthevulnerabilityoftheareassurroundingHerrer asCave. D ELINEATINGPROTECTIONAREASFORCAVESUSINGCONTAMINATIONVULNERABILITYMAPPINGTECHNIQUES:THECASEOF H ERRERI AS C AVE A STURIAS ,S PAIN 110 N JournalofCaveandKarstStudies, April2012

PAGE 113

Withrespecttothekarstnetwork(subfactorninthe COP + Kmethod),thecaveisaconduitinwhichthe piezometriclevelisvisibleinsomesectors.However,asthe caveitselfistheobjectofprotection,thewaterroute consideredissolelythatfromthesurfacetothewallsand roofofthecave.Oncethewaterorthepotential contaminantreachesthecavitywalls,theflowhasreached theobjective(Fig.8).Watertransitthroughthecavewas notconsideredforthepurposesofdelineatingthe protectionboundary. TheKvalueistheproductofthethreesubfactors describedabove,thet,r,andnsubfactors.Kvaluescanbe subdividedintothreeclasses,indicatingdifferentdegrees ofvulnerabilityofawatersourcetocontamination.This enablestheclassboundaryvaluestobemodifiedin responsetolegislationpassedtoprotectwatersupply sources(Fig.5).Forexample,accordingtotheSlovene approach(RavbarandGoldscheider,2007)vulnerabilityis highonlywhentheKvalueis1;inareaswherethetransit timeislessthan1day,thereisclearevidenceofanactive karstnetworkandthecontributionrateis 10%. However,foragivensource,accordingtothemethodology proposedbyAndreoetal.(2009),valuesof1to3indicatehighvulnerabilitytocontamination.Manycarbonate Figure7.SketchofthetracerteststoandwithinHerrer asCaveandtheflowvelocitiesdeduced. Figure8.ConceptualdiagramofthetransitofpossiblecontaminantstowardHerrer asCave. A.I.M ARI N ,B.A NDREO ,M.J IME NEZ -S A NCHEZ ,M.J.D OMI NGUEZ -C UESTA,AND M.M ELE NDEZ -A SENSIO JournalofCaveandKarstStudies, April2012 N 111

PAGE 114

aquiferspresentgoodpermeabilityasaresultofthe developmentofconduitsenablingrapidflowvelocitiesand directconnections.However,theseaquifersmightnothave beenidentifiedorexplored,andsoalowersubfactorn valuewouldnotbeawarded.Insuchacase,whichisvery similartothestudyareaconditions,theKfactorwould representahighlevelofvulnerabilityinaccordancewith theCOP + Kclassification,andamoderateoneaccording totheSloveneapproach. I NTRINSIC S OURCE V ULNERABILITY M AP ByconvertingtheCOPandKvaluesintotheir respectiveindicesandsummingthem(Fig.5),weobtain thesource(thecave)contaminationvulnerabilitymap.By directinterpretationofthevulnerabilityclasses,itis possibletofindtheareasthatneedvariousdegreesof protection(Andreoetal.,2009).Thevulnerabilitymapfor theHerrer asCaveobtainedusingtheCOP + Kmethod, classifyingtheKfactoraccordingtotheSloveneapproach (Kindex 5 1,TableXXIVinFig.5),showsalargezone classedashighlyvulnerable,coincidingwiththeareas wheretheresourcevulnerabilityisconsideredveryhigh (Fig.6).Theseareasarelocatedimmediatelysurrounding thecaveandinothersectorswithaquiferrechargevia influentstreams(Mar netal.,2010).Thevulnerabilityis consideredmoderateinareaswheretheresourcevulnerabilityishigh(COPindex 5 2)dueto,especially,slopesthat decreaseflowstowardsinfluentstreams.ThecaveÂ’s vulnerabilitytocontamination,evaluatedusingthe COP + Kmethod,withtheKfactorbeingscoredusing theCOP + Kmethod(Kindex 5 0,TableXXIIIinFig.5), isconsideredhighinalmosttheentirehydrogeological basin.Thevulnerabilityismoderateonlyinzoneswhere soilandQuaternarymaterialsofferacertaindegreeof protection,thezoneswhereresourcevulnerabilityis moderate(fig.6).Inanycase,theresultsobtained demonstratethatHerrer asCaveshowsahighlevelof vulnerabilitytocontaminationasaconsequenceofthe physicalcharacteristicsanddynamicsofthehydrogeologicalbasin. D ELINEATIONOFTHE P ROTECTION Z ONE Theultimateaimoftheprotectionzoningistoprotect thecaveagainstpossiblenegativeimpactsarisingfrom humanactivity.Thearrivalofcontaminantsubstances withinfiltrationwatermayprovokealterationswithinthe caveenvironment.Theentireareaconnectedtothecave viaflowsofsurfacewaterorgroundwatershouldbe consideredapossiblesiteofcontaminatingactivities,and therefore,shouldbeincludedwithinthedesignofthe protectionboundaryforthecave. IntheapplicationoftheCOP + Kmethodtoprotection zoning,theareastakenintoconsiderationwerethose classedashighvulnerability,requiringmaximumprotection,andmoderatevulnerability,requiringamoderate degreeofprotection.ThetwomapsproposedinFigure9 illustratetheneedforaprotectionboundaryforthiscave thatincludestheentirehydrologicalbasin,independentof thedegreeofinternalprotection. Forcomparison,Figure9showstheprotectionarea publishedintheBOPA(2008)andBOE(2010)together withtheoneproposedinthisstudy.Theareasthatwe proposehereasrequiringmaximumprotectionarelarger thanthosedefinedinBOPA(2008)andBOE(2010)and includepartofthewatershedthatfeedstheinfluent streams.ThisnetworkofinfluentstreamsonthenorthfacingslopesoftheSierradelCuerameritsspecial attentionbecauseofitshydrologicrelationwiththeaquifer inwhichthecaveislocated.Asexplainedabove,inthis area,surfacerunofftakesplace,especiallyoverthe sandstonesandquartziteofUnit2.Thissurfacerunoff infiltratesdownstreamintotheaquifer(Figs.2and3).The mapspresentedinFigure9showthatmostoftheriver bedsintowhichwaterinfiltratesarelocatedinthezone requiringmaximumprotection;therefore,anyhumanuse ofthewatershedshouldbecontrolledtoavoiddumpingor spillsthat,transportedbytherunoffwater,couldaffectthe cave.Accordingly,thehydrogeologicalbasinfedbythese riversisconsideredpartofthecaveÂ’sprotectionzoningina categorywetermwatershedtocontrol.Finally,unlikethe perimeterdefinedintheBOPAandBOE,wedonot consideritnecessarytoincludetheareaslocatedtothe northofthecaveandElBola uspringintheprotection zone.Thereasonisthatthewaterflowtowardthenorth preventsanycontaminantdepositedinthiszonefrom affectingthecave. C ONCLUSIONS Protectionzoningforcavesisapreventivemechanism ofgreatinterestfortheirconservation.Inthecaseofkarst aquifers,thisdelineationgoesbeyondthesimplemarking outofzonesrequiringdifferentdegreesofprotectioninthe immediatesurroundingsoftheitemtobeprotected,a spring,well,orcave.Thehydrogeologicalcharacteristicsof thesurroundingsshouldbeconsideredintheprocessby whichprotectionareasaredefined.Therefore,inthisstudy weadaptedtheCOP + Kmethod,whichisspecificfor groundwatercontaminationvulnerabilityassessmentin carbonateaquifers. Theresultsobtainedfromapplyingthismethodtothe Herrer asCaveshowthatprotectionmeasuresneedto considertheentiresectoroftheaquiferinwhichthecavity islocated.Suchprotectionzoningshouldincludethe riverbedsofinfluentstreamsandthecorresponding hydrographicwatersheds.IndependentlyoftheclassificationoftheKfactor,whichconcernsthekarstificationof thesaturatedzone,theextentoftheprotectedareas establishedrevealsthehighdegreeofvulnerabilityto contaminationofHerrer asCaveandtheinsufficiencyof thecurrentprotectionareadefinedforit.Land-use D ELINEATINGPROTECTIONAREASFORCAVESUSINGCONTAMINATIONVULNERABILITYMAPPINGTECHNIQUES:THECASEOF H ERRERI AS C AVE A STURIAS ,S PAIN 112 N JournalofCaveandKarstStudies, April2012

PAGE 115

Figure9.ProtectionareasproposedforHerrer asCaveonthebasisofthecontaminationvulnerabilityevaluationmadeusing theCOP + Kmethod(A)withtheKfactorbeingclassifiedaccordingtotheSloveneapproach(RavbarandGoldscheider,2007) and(B)withtheKfactorbeingclassifiedusingtheCOP + Kmethod(Andreoetal.,2009).Alsoshownistheprotectionarea recommendedinBOE(2010)andBOPA(2008). A.I.M ARI N ,B.A NDREO ,M.J IME NEZ -S A NCHEZ ,M.J.D OMI NGUEZ -C UESTA,AND M.M ELE NDEZ -A SENSIO JournalofCaveandKarstStudies, April2012 N 113

PAGE 116

planningisrequiredtoprotectthecave.Spainhasno nationallegislationdefiningtheactivitiesallowedin differentprotectedareas.However,potentiallypolluting land-useactivitiesshouldbeprohibitedorrestricted, especiallyactivitiescreatingmicrobialcontaminantsand typesofpollutingland-usepracticesthatgeneratedangerousleachates. Apartfromthefinalmappingobtainedofthestudy area,themaincontributionofthisstudyistheadaptation oftheCOP + Kmethodtodelineatecaveprotectionareas, especiallyforcavescontaininggroundwaterfloworclose toaspring.Theevaluationofacave’svulnerabilityto contamination,onthebasisofthecharacteristicsofthe physicalmediuminwhichitislocated,isausefultoolto ensurethecompatibilityofsurfacemanagementandcave conservation.Hydrogeologicalstudiesareessentialto achieveanadequateunderstandingoftheinteractions betweenthecaveanditssurroundings.TheCOP + K methodenablesustoplacethisknowledgeonacartographicbasis,whichconstitutesagoodinitialpointfor delineatingcaveprotectionareasandforpreventing possiblecontamination. Finally,itshouldbenotedthatprotectionareasarenot staticandneedtobeperiodicallyreviewedtodetermine whetherthephysicalvariablesjustifyingthedelineation remainunchanged.Ifenvironmentalalterationsaredetectedoriffurtherstudiesproducednewinformationabout hydrogeologicalrelationships,theproposedperimeter shouldbereviewedand,ifnecessary,modified,in accordancewiththenewreality. A CKNOWLEDGMENTS ThisresearchhasbeenpartlysupportedbytheRegional GovernmentofPrincipadodeAsturias(ContractCN-06177,UniversityofOviedo).Itisacontributiontoprojects CGL2008-06158BTEoftheSpanishMinistryofScience andHigherEducationandIGCP513ofUNESCO,andto ResearchGroupRNM-308fundedbytheRegional GovernmentofAndalusia(Spain).TheauthorsthankM. Field(Editor-in-Chief)andB.Schwartz(AssociateEditor), andanonymousreviewersfortheirconstructivecriticism. R EFERENCES Andreo,B.,Ravbar,N.,andV as,J.M.,2009,Sourcevulnerability mappingincarbonate(karst)aquifersbyextensionoftheCOP method:Applicationtopilotsites:HydrogeologyJournal,v.17, no.3,p.749–758,doi:10.1007/s10040-008-0391-1. Badiura,R.,andBrins ek,B.,1908.NovejameobCerknis kemjezeru.– Planinskivestnik,p.6–7,96–99,124–126,Ljubljana. BOE,1985,Ley16/1985,de25dejunio,delPatrimonioHisto rico Espan ol:Bolet nOficialdelEstado,no.155,p.20342–20352. BOE,2010,Decreto20/2010,de3demarzo,porelquesedelimita elentornodeproteccio ndelacuevadeLaHerrer a,enLaPereda, enelconcejodeLlanes:Bolet nOficialdelEstado,no.100, p.36592–36594. BOPA,2008,Resolucio nde27defebrerode2008,delaConsejer ade CulturayTurismo,porlaqueseapruebaelentornodeproteccio n provisionalparalacuevadeLaHerrer a,enlaPereda,concejode Llanes:Bolet nOficialdelPrincipadodeAsturias,v.74,7147p. Catani,F.,Fanti,R.,andMoretti,S.,2002,Geomorphologicrisk assessmentforculturalheritageconservation, in Allison,R.J.,ed., AppliedGeomorphology:TheoryandPractice:WestSussex,England,JohnWiley&Sons,p.317–334. Cuevas-Gonza lez,J.,Ferna ndez-Corte s,A.,Mun oz-Cervera,M.C., Andreu,J.M.,andCan averas,J.C.,2010,Influenceofdailyvisiting regimeintouristcaveatdifferentseasons, in Andreo,B.,Carrasco,F., Dura n,J.J.,andLaMoreaux,J.W.,eds.,AdvancesinResearchin KarstMedia,p.475–481. Dom nguez-Cuesta,M.J.,Jime nez-Sa nchez,M.,Rodr guez-Rodr guez, L.,Ballesteros,D.,Mele ndez,M.,Martos,E.,andGarc a-Sansegundo,J.,2010,Usodelageomorfolog ayelSIGparacaracterizarel impactodeactividadesminerasenzonaska rsticas:Elentornodela cuevadeLasHerrer as(Asturias,Espan a), in BerrezuetaAlvarado, E.,andDom nguezCuesta,M.J.,eds.,Te cnicasAplicadasala Caracterizacio nyAprovechamientodeRecursosGeolo gico-Mineros. VolumenI:DescripcionesMetodolo gica:Oviedo,Spain,Instituto Geolo gicoyMinerodeEspan a,p.80–90. Iriarte,E.,Sa nchez,M.A.,Foyo,A.,andTomillo,C.,2010,Geological riskassessmentforculturalheritageconservationinkarsticcaves: JournalofCulturalHeritage,v.11,no.3,p.250–258,doi:10.1016/ j.culher.2009.04.006. Jime nez-Sa nchez,M.,Dom nguez-Cuesta,M.J.,Garc a-Sansegundo,J., Stoll,H.,Gonza lez-Pumariega,P.,Fuente-Puente,G.,Mele ndez,M., Martos,E.,Vadillo,I.,Rodr guez-Rodr guez,L.,andAranburu,A., 2008a,Estudiopreliminardelageomorfolog adelacuevade Herrer asysuentorno(Llanes,Asturias,NoroestedeEspan a), in Benavente,J.,andGracia,F.J.,eds.,TrabajosdeGeomorfolog aen Espan a2006–2008:Ca diz,Spain,SociedadEspan oladeGeomorfolog aandUniversidaddeCa diz,p.45–48. Jime nez-Sa nchez,M.,Stoll,H.,Vadillo,I.,Lo pez-Chicano,M.,Dom nguez-Cuesta,M.,Mart n-Rosales,W.,andMele ndez-Asensio,M., 2008b,Groundwatercontaminationincaves:Fourcasestudiesin Spain:InternationalJournalofSpeleology,v.37,no.1,p.53–66. Jime nez-Sa nchez,M.,Vadillo,I.,Dom nguez-Cuesta,M.J.,Mele ndez, M.,Andreo,B.,Mar n,A.I.,Stoll,H.,Martos,E.,Gonza lezPumariega,P.,PotencianodelasHeras,A.,andGarc a-Sansegundo, J.,2010,EstudiodelkarstenlacuevadelasHerrer asysuentorno (Llanes,Asturias):Memoriafinaldeinvestigacio n2007–2010, Consejer adeCulturaPrincipadodeAsturias,179p. Jime nez-Sa nchez,M.,Dom nguez-Cuesta,M.J.,Aranburu,A.,and Martos,E.,2011,Quantitativeindexesbasedongeomorphologic features:Atoolforevaluatinghumanimpactonnaturalandcultural heritageincaves:JournalofCulturalHeritage,v.12,p.270–278, doi:10.1016/j.culher.2011.01.004. Jorda Cerda ,F.,andMalloViesca,M.,1972,Laspinturasdelacuevade LasHerrer as(Llanes,Asturias):RevistadelaFacultaddeFilolog a SeminariodePrehistoriayArqueolog adelaUniversidadde Salamanca,v.23,p.306–311. Kepa,T.,2001,KarstconservationinSlovenia:ActaCarsologica,v.30, no.1,p.143–164. Lera,T.,2009,TheVirginiaCaveProtectionAct:Areview(1966–2009): JournalofCaveandKarstStudies,v.71,no.3,p.204–209. Margat,J.,1968,Vulne rabilite desnappesd’eausouterrainea `lapollution: Basesdelacartographie:Orle ans,France,BureaudeRecherches Ge ologiquesetMinie res,Document68SGL198HYD. Mar n,A.I.,Andreo,B.,Jime nez-Sa nchez,M.,Dom nguez,M.J.,and Mele ndez,M.,2010,Delimitacio ndelper metrodeproteccio ndela cuevadeHerrer as(Llanes,Asturias)medianteelme todoCOP + K, in Proceedings,IIICongresoEspan olsobreCuevasTur sticas,CUEVATUR2010:Cuevas:patrimonio,naturaleza,cultura,yturismo: Aracena(Huelva).Spain,p.451–464. Marqu nez,J.,1989,S ntesiscartogra ficadelaRegio ndelCueraydelos PicosdeEuropa:TrabajosdeGeolog a,v.18,p.137–144. Ortun o,A.,Mele ndez,M.,andRodr guez,M.L.,2004,Relacio nentre litolog aycaracter sticashidroqu micasdelasaguassubterra neas: ReddeControldelaCalidaddelPrincipadodeAsturias:Bolet n Geolo gicoyMinero,v.115,no.1,p.35–46. Pulido-Bosch,A.,Mart n-Rosales,W.,Lo pez-Chicano,M.,Rodr guezNavarro,C.,andVallejos,A.,1997,Humanimpactinatouristkarstic cave(Aracena,Spain):EnvironmentalGeology,v.31,no.3–4, p.142–149,doi:10.1007/s002540050173. D ELINEATINGPROTECTIONAREASFORCAVESUSINGCONTAMINATIONVULNERABILITYMAPPINGTECHNIQUES:THECASEOF H ERRERI AS C AVE A STURIAS ,S PAIN 114 N JournalofCaveandKarstStudies, April2012

PAGE 117

Ravbar,N.,andGoldscheider,N.,2007,Proposedmethodologyof vulnerabilityandcontaminationriskmappingfortheprotectionof karstaquifersinSlovenia:ActaCarsologica,v.36,no.3,p.397–411. Sa nchez,M.A.,Foyo,A.,Tomillo,C.,and Iriarte,E.,2007,Geologicalrisk assessmentoftheareasurroundingA ltamiraCave:Aproposednatural RiskIndexandSafetyFactorforp rotectionofprehistoriccaves: EngineeringGeology,v.94,p.180–200,doi:10.1016/j.enggeo.2007.08.004. Sa nchez-Moral,S.,Can averas,J.C.,Soler,V.,Saiz-Jime nez,C.,Bedoya,J., andLario,J.,2002,Laconservacio ndelmonumento, in Lasheras,J.A., ed.,RedescubrirAltamira:Madrid,TunerEdiciones,p.245–257. SC.85/CONF.008/09,UnitedNationsEducational,ScientificandCultural Organization.9 th sessionoftheCommittee,Paris,France,2–6December 1985,09COMXA-Inscription:AltamiraCave(Spain),23p. V as,J.M.,Andreo,B.,Perles,M.J.,Carrasco,F.,Vadillo,I.,and Jime nez,P.,2006,Proposedmethodforgroundwatervulnerability mappingincarbonate(karstic)aquifers:theCOPmethod:ApplicationintwopilotsitesinSouthernSpain:HydrogeologyJournal,v.14, no.6,p.912–925,doi:10.1007/s10040-006-0023-6. WHC-08/32.COM/24Rev,2009,Decisionsadoptedatthe32ndsessionof theWorldHeritageCommittee,ConventionConcerningtheProtectionoftheWorldCulturalandNaturalHeritage.(QuebecCity,2008): Paris,UNESCOWorldHeritageCentre,230p. Zaporoz ec,A.,1994,Conceptofgroundwatervulnerability, in Vrba,J., andZaporoz ec,A.,eds.,Guidebookonmappinggroundwater vulnerability:Hanover,HeinzHeise,InternationalContributionsto Hydrogeology16,p.3–8. Zwahlen,F.,ed.,2004,COSTAction620.VulnerabilityandRisk MappingfortheProtectionofCarbonate(Karstic)Aquifers:Final Report.Brussels,EuropeanCommission,Directorate-generalXII Science,ResearchandDevelopment,297p. A.I.M ARI N ,B.A NDREO ,M.J IME NEZ -S A NCHEZ ,M.J.D OMI NGUEZ -C UESTA,AND M.M ELE NDEZ -A SENSIO JournalofCaveandKarstStudies, April2012 N 115

PAGE 118

MICROBIOLOGICALACTIVITIESINMOONMILK MONITOREDUSINGISOTHERMAL MICROCALORIMETRY(CAVEOFVERSCHEZLEBRANDT, NEUCHATEL,SWITZERLAND) O LIVIER B RAISSANT 1 *,S ASKIA B INDSCHEDLER 2 ,A LMA U.D ANIELS 1 ,E RIC P.V ERRECCHIA 2 AND G UILLAUME C AILLEAU 2 Abstract: Studiesoftheinfluenceofmicrobialcommunitiesoncalciumcarbonate depositsmostlyrelyonclassicalormolecularmicrobiology,isotopicanalyses,and microscopy.Usingthesetechniques,itisdifficulttoinfermicrobialactivitiesinsuch deposits.Inthiscontext,weusedisothermalmicrocalorimetry,asensitiveandnondestructivetool,tomeasuremicrobialactivitiesassociatedwithmoonmilk ex-situ .Upon theadditionofdilutedLBmediumandothercarbonsourcestofreshmoonmilksamples, weestimatedthenumberofcolonyformingunitspergramofmoonmilktobe4.8 3 10 5 6 0.2 3 10 5 .Thisnumberwasclosetotheclassicalplatecounts,butonecannotassume thatallactivecellsproducingmetabolicheatwereculturable.Usingasimilarapproach, weestimatedtheoverallgrowthrateandgenerationtimeofthemicrobialcommunity associatedwiththemoonmilkuponadditionofvariouscarbonsources.Therangeof apparentgrowthratesofthechemoheterotrophicmicrobialcommunityobservedwas between0.025and0.067h 2 1 andgenerationtimeswerebetween10and27hours.The highestgrowthrateswereobservedforcitrateanddilutedLBmedium,whilethehighest carbon-sourceconsumptionrateswereobservedforlowmolecularweightorganicacids (oxalateandacetate)andglycerol.Consideringtherapiddegradationoforganicacids, glucose,andothercarbonsourcesobservedinthemoonmilk,itisobviousthatupon additionofnutrientsduringsnowmeltingorrainfallthesecommunitiescanhavehigh overallactivitiescomparabletothoseobservedinsomesoils.Suchcommunitiescan influencethephysico-chemicalconditionsandparticipatedirectlyorindirectlytothe formationofmoonmilk. I NTRODUCTION Manybiogeochemicalprocessesinvolvingmanydifferentmicrobialactivitiestakeplaceincaves(seeBarton andNorthup,2007,andNorthupandLavoie,2001,for reviews).Mostofthesemicrobialprocessesareslowdueto theoligotrophicnatureofcaveenvironments(Mulec, 2008).Amongtheresultsoftheseprocesses,moonmilkis anubiquitouscavedepositcomposedofvariousmineral components,suchasneedle-fibercalcite(NFC)and nanofibers.Bothbioticandabioticmechanismshavebeen proposedforthedevelopmentofmoonmilk,andmultiple mechanismsarepossible,butmoonmilkisgenerally believedtobetheresultofmicrobialactivity(Can averas etal.,2006;Mulecetal.,2002).Thisdeposithasstrong similaritieswithsecondarycalciumcarbonateaccumulationsobservedinsoilsthataremainlycomposedofNFC andnanofibres.Manypapershaveprovidedstrong supportingevidenceforthebiogenicoriginofNFCin soils(Callotetal.,1985;Phillipsetal.,1987;Phillipsand Self,1987;Cailleauetal.,2009a,b;Curryetal.,2009). Moreover,recentinvestigationsemphasizethepotential implicationoforganictemplatesasaprecursorto mineralizednanofibres(Cailleauetal.,2009b;Bindschedler etal.,2010).However,thebiogenicoriginisstilldebated forbothmoonmilkandsoilsecondarycalciticdeposits composedofNFCandnanofibres(Lacelle,2010;Borsato etal.,2000;EngelandNorthup,2008).Inorderto investigatethebiogenicoriginofmoonmilk,manystudies haveusedelectronmicroscopy,bacterialcultures,orboth (Can averasetal.,2006;Mulecetal.,2002;Cailleauetal., 2009a,b;Bindschedleretal.,2010).Electronmicroscopyof naturalmoonmilksamplesprovidesusefulinformation onthespatialrelationshipbetweenmicroorganismsand calciumcarbonateneedlesandnanofibres,butitmustbe recognizedthatonlyalimitednumberofsamplescanbe observed,whichleadstoconclusionspotentiallyspeculative.Similarly,culture-basedmethodsprovidegreatinsight intopotentiallyinvolvedmicrobialprocesses.Forexample, culturemethodswereusedtoshowtheabilityofsoil bacteriatoprecipitatecalciumcarbonate(Boquetetal., *CorrespondingAuthor:olivier.braissant@unibas.ch 1 LaboratoryofBiomechanics&Biocalorimetry,c/oBiozentrum/Pharmazentrum, UniversityofBasel,Klingelbergstrasse50-70,CH-4056Basel,Switzerland 2 InstituteofGeologyandPaleontology,UniversityofLausanne,QuartierUNILDorigny,Ba ˆtimentAnthropole,CH-1015Lausanne,Switzerland O.Braissant,S.Bindschedler,A.U.Daniels,E.P.Verrecchia,andG.Cailleau–Microbiologicalactivitiesinmoonmilkmonitoredusing isothermalmicrocalorimetry(CaveofVerschezleBrandt,Neuchatel,Switzerland). JournalofCaveandKarstStudies, v.74,no.1, p.116–126.DOI:10.4311/2011JCKS0213 116 N JournalofCaveandKarstStudies, April2012

PAGE 119

1973).However,alargepartofthemicrobialpopulation cannotbecultivated,andtherefore,arenottakeninto account(Ammanetal.,1995).Finally,nucleicacid-based studiesprovideinsightonthemicrobialcommunity,but itisoftendifficulttoinferpopulationsizesandspecific activitiesfrommoleculardataunlessoneusesstable isotopeprobing(Konhauser,2007). Inthiscontext,isothermalmicrocalorimetrycanprovideanothervaluabletoolforobtainingdataonmicrobial activitiesassociatedwithmoonmilk.Isothermalmicrocalorimetryisanefficientmeansofmonitoringmicrobial activitiesthroughmeasurementofheatproductionrates. Isothermalmicrocalorimetryisextremelysensitive,and heatflowsaslowas20to200nWaresufficienttoproduce areliablesignal(Braissantetal.,2010;Wadso ¨,2002). Assumingthatatypicalsinglebacterialcellproducesabout 2pWwhenactive(Higuera-Guissetetal.,2005;James, 1987),only10,000to100,000bacteriapersample(i.e., per3or4mlmicrocalorimetricampoule)arerequiredto produceadetectablesignalinmostcommercialisothermal microcalorimeters.Sinceonlyheatproductionismeasured, isothermalmicrocalorimetryisnon-destructiveandnoninvasive,allowingsamplestobestudied ex-situ without disturbanceotherthanplacingtheminthecalorimetric vessel.Althoughthesearestillsignificantinteractions,they canbeconsideredminimalcomparedtoothertechniques. Duetotheseadvantages,isothermalmicrocalorimetryhas beenwidelyusedforsoils(Rongetal.,2007;Wadso ¨,2009), butithadnotyetbeenappliedtocaves. Inthisstudy,weinvestigatedthemicrobialactivitiesin moonmilksupplementedwithdifferentcarbonsources. Fromthedata,thenumberofactivebacteriaandthe varioussubstratedegradationratesareinferred.This approachprovidesavaluablesetofdatathatcomplement previousstudiesonmoonmilkandbiomineralization. M ATERIALSAND M ETHODS G EOLOGICAL S ETTINGS VerschezleBrandt(46 u 56 9 16 0 N,6 u 28 9 22 0 E)isa limestonecave(Sequanianstage)locatedintheSwissJura mountains4kmnorthofthevillageofLesVerrie `res (Fig.1A)andonly1200metersfromtheFrenchborder.The caveis63metersdeepand336meterslong(Fig.1B)andhas beenpreviouslydescribed(Gigonetal.,1976).Theaverage temperatureinthecaveisabout10 u C(Perrin,2003). S AMPLE C OLLECTION Moonmilksamplesfromthelocationindicatedin Figure1BwerecollectedasepticallyonJuly17,2009. Althoughmoonmilkcanbefoundinmostofthecave,this locationwaschosenbecauseitislessaccessibleandthus lesscontaminated.Allthematerialusedwassterile,and sampleswereplacedintosterilesamplingbagsorsterile 50mlpolypropylenetubes.Approximately100gofslightly wetmaterialwascollected.Sampleswerestoredinacooler containingblueice,andlaterat4 u Cuntilanalysis. Samplescollectedformicrocalorimetrywereprocessedand incubatedinthemicrocalorimeterwithintwentyhours aftercollection.Additionalmoonmilksampleswere collectedformicroscopicobservationsandwerestoredat 4 u Cforoneweekbeforepreparationandobservation. Thesesamplesweretakenfromspecificmoonmilkpatches selectedfortheirdarkercolor,whichweassumedtohave slightlyhigheramountsoforganicmatter(authors’ personalobservations;noorganicmattermeasurements wereperformedinthiscave).Organicmatterislikely broughttothecavethroughrockfractures(Perrin2003). Aswedidnotexpecttofindalotofmicrobesin oligotrophicmoonmilkdeposits,sothesespotswithputativelymoreorganicmatterwereconsideredtopotentially haveagreaternumberofmicrobes. M ICROCALORIMETRIC A NALYSIS Moonmilksampleswerepreparedasdescribedbelow andintroducedintothemicrocalorimeter(TAM48, Waters/TA,Delaware)andkeptintheequilibration position,insidethemicrocalorimeterbutnotincontact withthethermopilethatisthesensingelement,for 15minutestoensurepreliminarythermalequilibration. Then,sampleswereloweredintothemeasuringpositionin contactwiththethermopile.Sufficientthermalequilibrationwasachievedafter45minutes,andheatflow measurementsstartedapproximatelyonehourafter insertionintothemicrocalorimeterandlastedforabout 5days.Allsamples,controls,andblanksweremeasuredin triplicateunlessstatedotherwise.Allisothermalmicrocalorimetrymeasurementswereperformedat25 u Cbecause ofthetechnicallimitationsofourmicrocalorimeter,which cannotbeoperatedatcavetemperatures(mostcalorimetersoperateatroomtemperatureorabove).Thereforea temperatureclosetothatusedinmanyotherstudieswas chosen(Laizetal.,2000). E STIMATIONOFTHE N UMBEROF M ICROBIAL C ELLS P RESENTINTHE M OONMILK Samplesofmoonmilkwerecoarselygroundand homogenizedinasterilePetridishusingasterilescalpel blade.Toestimatethenumberofactivecellsinthe moonmilk,ampoulesfilledwith2gofthiscoarselyground moonmilkwerepreparedandsupplementedwith800 m lof 50 3 dilutedLB(Difco–tryptone0.2gL 2 1 ,yeastextract 0.1gL 2 1 ,NaCl0.1gL 2 1 )asthecarbonsourcesavailable inthismediumwillsupportgrowthofmanychemoheterotrophicmicroorganisms.Followingtheadditionof medium,thepreparedampoulesweresealedandmicrocalorimetricmeasurementswereperformedat25 u Cas describedabove.Inadditiontothemicrocalorimetric measurements,platecountswereperformedon50 3 diluted LAmedium(Difco–tryptone0.2gL 2 1 ,yeastextract 0.1gL 2 1 ,NaCl0.1gL 2 1 ,agar15gL 2 1 )andon10 3 dilutedR2Amedium(BactoPeptone0.05gL 2 1 ,Bacto O.B RAISSANT ,S.B INDSCHEDLER ,A.U.D ANIELS ,E.P.V ERRECCHIA,AND G.C AILLEAU JournalofCaveandKarstStudies, April2012 N 117

PAGE 120

yeastextract0.05gL 2 1 ,Bactocasaminoacids0.05gL 2 1 glucose(Fluka)0.05gL 2 1 ,solublestarch(Fluka) 0.05gL 2 1 ,KH 2 PO 4 (Fluka)0.03gL 2 1 ,MgSO 4 N 7H 2 O (Fluka)0.005gL 2 1 ,Bactoagar15gL 2 1 ).Dilutedmedia havepreviouslybeenshowntobeappropriateforplate countsofmicroorganismsfromoligotrophicwater(Kawai etal.,1999;Janssenetal.,2002;Bull,2004;Segawaetal., 2011).Allagarplateswereincubatedat25 u Cforthesame timeperiodasthemicrocalorimetricmeasurements(i.e., 5days)toobtainresultscomparabletothoseobtained usingmicrocalorimetry.Sterilitycontrolswereperformed usinguninoculatedplatesforeachmedium. Asimilarexperimentwasperformedtoestablish controlsallowingquantificationofthenumberofcells presentinthemoonmilk. Bacillussubtilis waschosenasthe controlbecauseofitsabilitytoconsumeawiderangeof carbonsourcesanditsstructuralandmetabolicsimilarities withtheactinobacteriacommonlyfoundincaves.In addition, B.subtilis iseasiertohandle,comparedto filamentousactinobacteria.Ampouleswerefilledwith2g ofcoarselygroundmoonmilkandautoclavedtwiceat5to 6hoursintervals.Theseampouleswereinoculatedwitha serialdilutionofasuspensionof Bacillussubtilis (NEU16 –Neucha ˆtelUniversityculturecollection)in50 3 diluted LB.The B.subtilis suspensionwaspreparedbycentrifugationofanovernightculturegrownin10 3 dilutedLB. Thepelletwasthenresuspendedin50 3 dilutedLB.This operationwasrepeatedtwicetoensurethatonly50 3 dilutedLBremained.Finallythepreparedampouleswere sealedandmicrocalorimetricmeasurementswereperformedat25 u Casdescribedabove.Blankswereprepared similarly,exceptnocellswereadded.Platecountswere performedonLAmedium(Difco–tryptone10.0gL 2 1 yeastextract5.0gL 2 1 ,NaCl10.0gL 2 1 ,agar15gL 2 1 )to estimatethenumberofcolonyformingunitsintroduced intotheampoules. M EASUREMENTOF M ICROBIAL A CTIVITIESIN M OONMILK Twogramsofcoarselyground-upmoonmilkwere placedinacalorimetricampouleandsupplementedwith 800 m lofchosencarbonsources.Unlessstatedotherwise, allthecarbonsourcesandantibioticswereobtainedfrom Fluka(Switzerland).Thecarbonsourcesat0.3mgmL 2 1 wereglucose,glycerol,mannitol,Na-oxalate,Na-citrate, Na-acetate,xanthan,starch,humicacids(saturatedsolution),50 3 dilutedLB(Difco),and50 3 dilutedLB supplementedwithafinalconcentrationof50 m gmL 2 1 ofchloramphenicol.Allcarbonsourcesanddehydrated mediawerepreparedindeionizedwater.Controlswere performedusingautoclaveddeionizedwater,andblanks wereperformedusingmoonmilkautoclavedtwice(25minutesat121 u Cat5to6hoursintervals)supplementedwith sterile50 3 dilutedLB.Preliminarytestshaveshown that800 m lofsolutionissufficienttowetthesample homogenouslywithoutaccumulationofsolutionatthe bottomofthesamples.Finally,samplesweresealedand introducedintheisothermalmicrocalorimeter. Apparentmicrobialgrowthrateandgenerationtime werecalculatedforthedifferentcarbonsourcesaccording tothemethodpreviouslydescribedinKimuraand Takahashi(1985)andinBarrosetal.(1999). S CANNING E LECTRON M ICROSCOPY (SEM) Sampleswerekeptat4 u Cduringoneweekpriorto SEMobservations.Observationswereperformedusinga TescanMiraLMUscanningelectronmicroscope.Samples werefixedwithosmiumtetroxidevapors,freeze-dried,and coatedwithgold(18nm)andcarbon(5nm).Observations wereperformedatadistanceof10mmandat10or15kV acceleratingvoltage,dependingonchargingartifacts. R ESULTS E STIMATIONOFTHE N UMBEROF M ICROBIAL C ELLS P RESENTINTHE M OONMILK Byusing Bacillussubtilis ,atypicalsoilbacterium,in autoclavedcalorimetricampoulescontainingsterilized moonmilk,undertheexperimentalconditionsdescribed above,weestimatedthatonecolony-formingunitgenerFigure1.A.Switzerlandmapindicatingthelocationofthe VerschezleBrandtcave(star)andSwissmaincities(dots). Notethatthecaveislocatedonly1200mfromtheFrench border.B.Horizontalprofileofthecave.Thearrowindicates thesamplingsite.Greyshadedareasindicatethepresenceof largeboulders. M ICROBIOLOGICALACTIVITIESINMOONMILKMONI TOREDUSINGISOTHERMALMICROCALORIMETRY (C AVEOF V ERSCHEZLE B RANDT N EUCHATEL ,S WITZERLAND ) 118 N JournalofCaveandKarstStudies, April2012

PAGE 121

atedaheatflowof3.0pW(Fig.2).Thisvalueisbasedon thefirsthourofrecordeddata.Duringthistime,changesin heatflowwerelessthan5%oftheinitialvalue;therefore wemadetheassumptionthatthenumberofCFUdidnot significantlyincrease.Itisalsoinaccordancewiththe rangegivenforbacteriabyJames(1987).Acleartime delayuntilmaximumactivityindicatedbytheheatflowis visiblewithdecreasingnumberofcells(Fig.2).Similarly, aslightdecreaseofmaximumactivitywithdecreasing inoculumisalsoobserved(Fig.2).Itislikelythatthe bacteriaatlowerdensitiesdidnothaveaccesstoallthe substrateforgeometricreasons.Inaddition,recyclingof partoftheintroducedbiomassmightexplainthehigher heatflowobservedathighdensities.Therefore,bothof thesefactors,restrainedaccesstocarbonsourcesand biomassrecycling,arelikelytoexplaintheobserved decreasingtrend.However,thetotalheatproducedafter 100hours(i.e.,theintegraloftheheatflowcurveover 100hours)wasmostlyconstant,withanaverageof2.2 6 0.3J(n 5 18),showingthatsubstrateconsumptionwas essentiallysimilarinallthesamples. Uponadditionof50 3 dilutedLBmediumtofresh moonmilksamples,weestimatedthattheamountofactive chemoheterotrophiccellspergramofmoonmilkfromVers chezleBrandtcaveis4.8 3 10 5 6 0.2 3 10 5 activemicrobial CFU(n 5 5).Thisestimationisofcoursedependantonthe estimationofthethermalpowerofoneCFU. Platecountson50 3 dilutedLAand10 3 dilutedR2A mediaincubatedover5daysshowcountsofthesameorder ofmagnitude(7.7 3 10 5 6 2.4 3 10 5 and7.8 3 10 5 6 2.9 3 10 5 CFUpergramofmoonmilk,respectively).Although higher,thesevaluesareinthesamerangecomparedtothe microcalorimetricdata.Itisnoteworthythattheproportionofactinobacteriadeterminedmicroscopicallyonthese plateswere81% 6 19%on50 3 dilutedLBand94% 6 6%on 10 3 dilutedR2A.Thisobservationisinagreementwith previousreportoftheassociationofactinobacteriawith moonmilk(Can averasetal.,2006),SEMobservations,and thestrongoxalate-degradingactivitiesmeasured(see below). M ICROBIAL A CTIVITIESIN M OONMILK Additionofthevariouscarbonsourcesresultedinan increaseinheatflow,indicatinganincreaseinmicrobial activity.Thisincreaseisusuallyobservedafteralagphase upto30hours.Afterreachingamaximum,theheatflow decreasesbacktovaluesclosetotheinitialvalues.The resultingheatflowpatternischaracterizedbyasingle peak,usuallyshowingasmallshoulder(Fig.3),exceptin thecaseof50 3 dilutedLB,wheretwoclearlyseparated peaksareobserved(Fig.3A). Strongdifferencesareobservedbetweenheatflow patternsresultingfromtheadditionofthevariouscarbon sources(Fig.3).Adeterminationoftheapparentgrowth rateandgenerationtime(Table1)clearlyshowsthe differencesinmicrobialstimulationgeneratedbythe differentcarbonsources.Intheconditionsoftheexperiment,citrate,dilutedLBmedium,xanthan,glucose,starch, Figure2.Heatflowpatterngeneratedbythegrowthofserial10-folddilutionsin50 3 dilutedLBof Bacillussubtilis addedto sterilizedmoonmilk.Insidepanelshowstherelationshipbetweenthenumberofcellsandtheheatflowgeneratedduringthe firsthour,allowinganestimateofthethermalpowerofonecolonyformingunit.Thecurveforthemostcellsistheonewiththe left-mostpeak. O.B RAISSANT ,S.B INDSCHEDLER ,A.U.D ANIELS ,E.P.V ERRECCHIA,AND G.C AILLEAU JournalofCaveandKarstStudies, April2012 N 119

PAGE 122

andglycerolsupportedahigherapparentgrowthratethan humicacids,mannitol,acetate,andoxalate.Deionizedwater additionresultedinverylittleheatflowincrease,possiblydue tothedissolutionofcarbonatemineralsordesorption processesreleasingaverysmallamountofnutrients.Asmall riseintheheatflowwasnotobservedwhenautoclaved moonmilkwassupplementedwitheitherdilutedLBor deionizedwater,emphasizingthefactthatitismostlikely relatedtoabiologicalprocess.Ontheotherhand, consumptionratesshowadifferentpicture.Themaximum consumptionrateswerecalculatedassumingthechemical equationslistedinTable2.Anobvioustrendappearswhen plottingGibbsfreeenergyavailableversusconsumptionrate (Fig.4; r 5 0.95, n 5 6, p 0.05).Thehighestcalculated consumptionrateisgivenbytheoxalatesubstrateandisin agreementwiththehighnumbersofactinobacteriaobserved intheplatecounts.Thisgroupofbacteriaisknownforits greatabilitytodegradeoxalate(Sahin,2003;2004). Amuchlongerlag(about20hourslonger)isobserved afteradditionof50 3 dilutedLBcontainingchloramphenicol(Fig.3).Nevertheless,theresultingheatflowpattern lookssimilar,althoughcontractedintime.Thisresult demonstratesthataratherlargeportionofthemicrobial communityissensitivetochloramphenicolandmayshow thatthereislittlefungalactivity,becausefungishouldnot besensitivetosuchchloramphenicolconcentration. E LECTRON M ICROSCOPY Electronmicroscopyofthemoonmilksamplesrevealed thepresenceofnumerousmorphotypesofmicroorganisms. Mostobservationsshowfilamentousmicroorganisms. Accordingtotheirmorphology,thesemicroorganisms couldmostlikelyberelatedtoactinobacteriaandtheir arthrospores(Fig.5B,C).Thisisinagreementwithour observationsonagarplates.Otherfilamentousorganisms, showingresemblancetohyphomicrobiaceae,arealso observed(Fig.5D).Finally,largerfilaments,branchedor not,areinterpretedasfungi(Fig.5E,F).Mostmicroorganismswereobservedatthesurfaceofaggregatesof intactsamplesofmoonmilkaggregates.Incontrast, observationsperformedoncutorcoarselygroundmoonmilksamplesshowfewermicroorganisms.Consideringthe slowgrowthrateobserved,evenundertheadditionof carbonsources,andtheslowappearanceofcoloniesin thelaboratory,itislikelythatthebacterialmorphotypesobservedherebelongtooursample.However,the possibilityofsomegrowthduringthestorageofthe samplescannotbecompletelydismissed. Figure3.Representativeheatflowpatternsgeneratedbytheadditionofvariouscarbonsourcestomoonmilk.A.Moonmilk samplessupplementedwith50 3 dilutedLBmediumand50 3 dilutedLBmediumsupplementedwithchloramphenicoland deionizedwater.B.Moonmilksupplementedwithglucoseandpoly-alcohols.C.Moonmilksupplementedwithorganicacids.D. Moonmilksupplementedwithpolymers. M ICROBIOLOGICALACTIVITIESINMOONMILKMONI TOREDUSINGISOTHERMALMICROCALORIMETRY (C AVEOF V ERSCHEZLE B RANDT N EUCHATEL ,S WITZERLAND ) 120 N JournalofCaveandKarstStudies, April2012

PAGE 123

D ISCUSSION Ourmicrocalorimetricdataclearlyshowthatmicrobial activitycaneasilyandaccuratelybemeasuredinthe moonmilksamplesduring ex-situ analyses.Inaddition,the numbersofcolony-formingunitsofcultivablebacteria estimatedbyplatecountsafter5daysandbymicrocalorimetryareinagreement.However,therelationship betweentheplatecountandthemicrocalorimetricestimatesshouldbeconsideredwithcare,sinceonecannotbe surethatallthebacteriathatproducedmetabolicheat uponadditionof50 3 dilutedLBarecultivable.Viablebut notculturablecellsmighthavecontributedtotheheat production.Indeed,after32daysofincubation,numbers ofculturablebacteriameasuredontheagarplates increasedto9 3 10 6 CFUpergramofmoonmilk,and showednofurtherincrease(datanotshown).Therefore, themeasured4.8 3 10 5 activeCFUpergramofmoonmilk wouldrepresentonly5.3%ofthetotalnumberofcultivable bacteriaandanevensmallerfractionofthetotalbacteria, includingthosenotculturable.Inaddition,thediscrepancy betweenmicrocalorimetricestimatesandplatecountsat5 and32dayssuggeststhataratherlargeproportionofthe bacteriamightbedormant.Thisdormantfractioncould beactinobacteriaarthrospores(GoodfellowandWilliams, 1983),butcouldalsoincludedormantcells sensustricto withnovisibleactivity(Costertonetal.,1995).These dormantcellsarecommonatverylowgrowthrates(Pirt 1987)andmaycontributetothemaintenanceofmicrobial diversityinoligotrophicenvironments(JonesandLennon 2010).Inoligotrophicenvironmentssuchasacave,lackof nutrientsislikelytobethemaincauseofdormantornonculturablestates(ColwellandGrimes,2000). Thesediscrepanciesemphasizethelimitationsofisothermalmicrocalorimetry.Indeedmicrocalorimetryisa blindtoolthatonlymeasuresheatproductionrate;inthis studymetabolicheatproductionrate.Togetabetter pictureofmicrobialprocessesinmoonmilk,microcalorimetryshouldbecoupledwithothertools,allowingan estimateofthemicrobialpopulationsize.Inaddition,the compositionofthecommunity,determinedbymolecular Table1.Apparentgrowthrateofthemicrobialcommunity whenexposedtodifferentsubstrates.Thegrowthrateand generationtimeweredeterminedusingisothermalmicrocalorimetrydataforthedifferentcarbonsourcesaccordingto themethodpreviouslydescribedinKimuraandTakahashi (1985)andinBarrosetal.(1999). CarbonsourceGrowthrate,h 2 1 Generationtime,h Oxalate0.0253 6 0.000327.4 6 0.3 Acetate0.0332 6 0.003721.0 6 2.4 Glycerol0.0456 6 0.004115.3 6 1.4 Citrate0.0668 6 0.007510.4 6 1.2 Glucose0.0506 6 0.007313.9 6 2.0 Mannitol0.0336 6 0.000920.7 6 0.6 Xanthan0.0565 6 0.003412.3 6 0.7 Starch0.0474 6 0.002514.7 6 0.8 Humicacids0.0369 6 0.000818.8 6 0.4 LB50x0.0579 6 0.005112.0 6 1.1 Figure4.RelationshipbetweenGibbsfreeenergyofthe substratesÂ’oxidationandtheirmaximumconsumptionrate. Table2.Consumptionratesofthevariouscarbonsourcescalculatedbasedontheassumptionthatonlyaerobicrespiration occurred.Equationsarelistedwiththethermodynamicparametersusedtoconvertmaximumheatflow(WorJs 2 1 )into substrateconsumptionrate(nmolesh 2 1 g 2 1 ). Equation Substrate commonname D G 0 9 (KJmole 2 1 ) Maximumheat flow( m W) Substratemaximum consumptionrate (nmolesh 2 1 g 2 1 ) a 2C 2 H 2 O 4 + O 2 R 4CO 2 + 2H 2 Ooxalate 2 32818 6 298 6 11 C 2 H 4 O 2 + 2O 2 R 2CO 2 + 2H 2 Oacetate 2 87234 6 1071 6 21 2C 3 H 8 O 3 + 7O 2 R 6CO 2 + 8H 2 Oglycerol 2 165370 6 476 6 5 2C 6 H 8 O 7 + 9O 2 R 12CO 2 + 8H 2 Ocitrate 2 214453 6 145 6 1 C 6 H 12 O 6 + 6O 2 R 6CO 2 + 6H 2 Oglucose 2 286057 6 136 6 1 2C 6 H 14 O 6 + 13O 2 R 12CO 2 + 14H 2 Omannitol 2 308149 6 729 6 4 (C 6 H 12 O 6 ) n + 6O 2 R 6CO 2 + 6H 2 O b starch 2 2860 b 31 6 2 b 19 6 2 b a Assumingthatonlythereactionconsideredtakesplace. b Convertedintoglucoseequivalents. O.B RAISSANT ,S.B INDSCHEDLER ,A.U.D ANIELS ,E.P.V ERRECCHIA,AND G.C AILLEAU JournalofCaveandKarstStudies, April2012 N 121

PAGE 124

Figure5.Scanningelectronmicrographsofmoonmilkaggregatesandmicroorganismsfoundinthem.A.Generalviewofthe moonmilksampleshowingmostlymonocrystallineandserrated-edgedneedle-fibercalcite.B.Growingfilamentous microorganismsresemblingactinobacteria.C.Chainsofcellsresemblingactinobacteriaarthrospores.Notethatspores arrangedinspiralsareusuallyassociatedtothegenus Streptomyces (seeHoltetal.,1994).D.Chainofappendagedbacterium. E.andF.Fungalfilaments. M ICROBIOLOGICALACTIVITIESINMOONMILKMONI TOREDUSINGISOTHERMALMICROCALORIMETRY (C AVEOF V ERSCHEZLE B RANDT N EUCHATEL ,S WITZERLAND ) 122 N JournalofCaveandKarstStudies, April2012

PAGE 125

methods,wouldallowrelatingtheactivitiesmeasuredto themembersofthiscommunity.However,inourstudy, theuseofaspecificantibioticallowsustopointtopossibly importantmembersofthemoonmilkpopulation.Upon additionofchloramphenicol,thepeakofactivitywas stronglydelayed,butnotsuppressed.Undoubtedly,apart ofthebacterialcommunitypresentwassensitivetothis antibiotic.Sincefungishouldnotbeaffectedbysuchlow concentrationsofchloramphenicol,theymightberesponsibleforapartoftheremainingactivity.Also,considering thewideoccurrenceofchloramphenicolresistanceamong actinobacteria(ShawandHopwood,1976)andtheirhigh proportioninoursamples,itseemsthatactinobacteria mightbeakeyplayerinthiscommunity.RecentpublicationshaveemphasizedtheimportanceofActinobacteriaincaves.Forexample,Laizetal.(2000)usingboth culturalandmolecularmethodstostudythemicrobial communityonstalactites,foundthatabout50%ofthe isolatedstrainswereactinomycetesandthatthemost prominentTGGEbandwasmostlikelyrelatedtoaclose relativeof Nocardiopsis, anotheractinobacteria.Similarly, 16%ofaclonelibraryobtainedfromthewallsofakarstic caveinSloveniawereactinobacteria(Pas ic etal.,2009), makingactinobacteriathesecondmostabundantgroup aftergamma-proteobacteria(about33%).Finally,combiningmolecularDNAandRNAbasedfingerprinting, targetingthetotalcommunityandthemetabolicallyactive partofthecommunity,respectively,Portilloetal.(2008, 2011)haveshownthatmoonmilkcontainedasubsetofthe communitypresentinthewhitecolonizationfromthe AltamiraCaveandthatactinobacteria,amongothers,were animportantpartofthemetabolicallyactivemicrobial population.Theauthorsalsoreportthatabout20%of thecommunitydetectedthroughDNA-basedfingerprintingremainsundetectedthroughRNA-basedfingerprinting,emphasizingthataratherlargeproportionofthe communityshowsundetectablemetabolicactivity.Furthermore,thenanorespirometrymeasurementsperformed inthestudyperformedbyPortilloetal.(2011)indicatethat oncewhitecolonizationsarecalcifiedintomoonmilk,the metabolicactivitydecreasesbyaboutafactorof20. Themicrocalorimetricdataobtainedduringthisstudy alsoshowthatitispossibletofurthercharacterizethe metabolicallyactivepartofthecommunitybysupplementingthemoonmilksampleswithdifferentcarbonsources, indicatingtheresponsesoftheactivepartofthemicrobial communitytospecificcarbonsources.Suchresponseis expressedasthegrowthrateofthecommunityinresponse toacarbonsourceorasthemaximumconsumptionrateof thiscarbonsource.Inourstudy,thecalculatedapparent growthratesarecomparabletogrowthratesobservedin soils(Barrosetal.,1995;1999;Parinkinaetal.,1973). However,wehavetorecognizethatthetemperaturein themicrocalorimeterwashigherthanthecaveenvironment,25 u Cinsteadof10 u C,whichmighthaveledtooverestimatingthesegrowthrates .Nevertheless,ourdatashow cleardifferencesinmicrobialgrowthratesobserveduponthe additionofthedifferentcarbonsources. Similarly,substrateconsumptionratesoforganicacids andglucosecalculatedinthisstudyarewithintherangeof thoseobservedinforestsoils(Fujiietal.,2010;vanHees etal.,2002).Bothoftheseobservationssuggestthat,inthis cave,themicrobialcommunitycanbeasactiveasinsoils. Onemajorlimitationinsuchacommunityisthesubstrate inputthatismainlylinkedtotheleachingofmaterialfrom thesoilabovethecave.Suchleachingcanonlyoccur duringrainorstormeventsandspringsnowmelt. Therefore,theactivity,intheabsenceofadditionof substrate,isextremelylowinthisnutrient-deficientsystem (Mulec,2008).Highmetabolicandheat-productionrates canbemeasuredincaveswhenthesampleiscollectedfrom microbialmats(Rohwerderetal.,2003).Althoughorganic acidandglucosehavereceivedalotofattentionwith respecttotheirdegradationrateintheenvironment,other carbonsources,suchasmannitol,starch,xanthan,or humicacid,havebeenneglectedintheliterature.Therefore,wecannotcomparetheirmeasureddegradationand associatedapparentgrowthratesfromourstudywithdata fromotherhypogeanenvironmentsorsoils. Withrespecttotheformationofmoon milk,theseresults pointtometabolicactivitiescapableofcreatingconditions favorablefortheprecipitationofcalciumcarbonate.Physicochemicalcharacterizationofthewaterpercolatinginthiscave duringrainfallhasshownthatitisundersaturatedwith respecttocalcite(Perrin,2003),andconsequently,ableto dissolvecarbonateminerals.Similarly,thesamestudyhas shownthatthispercolatingwaterischaracterizedbyan increasingcontentoftotalorganiccarbon.Inthiscontext,it canbeassumedthatthemicrobialconsumptionofthis organiccarbonmightprovidesufficientcarbonateionsand alkalinitytopromotecalciumcarbonateformation.Indeed, thegrowthofcavemicrobialpopulationsinadilutedmedium suchasnutrientbrothwasshowntoincreasethepHtovalues closetoorabovecarbonatemineralstability(i.e.,pH 5 8.4; PortilloandGonzalez,2011).Microbialmetabolismof organicacidsimportedinthesystemthroughthepercolating waterwillcontributetotheincreaseinpHsincemanyorganic acidshavehigherpHsthancarbonicacid.ThisincreaseinpH haspreviouslybeendemonstratedforoxalate(Braissantet al.,2002,2004).Similarly,Curryetal.(2009)have demonstratedthatuseofcalciumsuccinatebybacteria isolatedfromcottonballs(aspecifictypeofmoonmilk) yieldedasignificantproductionofcalciumcarbonate. Turnoverofhumicacidsorexopolymericsubstancessuch asxanthancouldalsofavortheprecipitationofcarbonate mineralsinthesameway.Inaddition,suchprocesseswillalso favorprecipitationofcarbonatemineralsbyreleasingthe calciumboundtotheexopolymers(Braissantetal.,2007; 2009;Duprazetal.,2009)orthehumicacidsandbyremoving precipitationinhibitors. Thelargeproportionofactinobacteriaobservedinthis study,aswellasinothers(Laizetal.,2000;Stomeoetal., O.B RAISSANT ,S.B INDSCHEDLER ,A.U.D ANIELS ,E.P.V ERRECCHIA,AND G.C AILLEAU JournalofCaveandKarstStudies, April2012 N 123

PAGE 126

2008;Cuezvaetal.,2009),andtheirstrongabilitytodegrade oxalate(Sahin2003,2004),exopolymericsubstances(Qian etal.,2007;Muchova etal.,2009;Tanetal.,2009;Gaskellet al.,2010),andhumicacid(Badisetal.,2009;Darietal., 1995)emphasizeoncemoretheirpotentialroleinthe formationofmoonmilkbycreatingthenecessaryphysicochemicalconditionsforcarbonate-mineralprecipitation. Theroleofactinobacteriaintheformationofmoonmilkwas previouslydiscussedbyCan averasetal.(2006).According totheirproposedmodel,theformationofamicrobial biofilmwouldleadtoastep-by-stepformationofmoonmilk usingbacterialand/orfungalhyphaeasthetemplate. Filamentousactinobacteriawereshowntoproduceneedleshapedcrystalsincultures,buttheneedlesresultingfrom suchcultureshavedifferentsizesandshapesthanthose observedinmoonmilk(Fig.6).Therefore,anotherprocess isthoughttobeattheoriginofthecalciticfeaturesobserved inthecaveVerschezleBrandt.Nobiofilmhasbeen observedinthesamplescollectedinthatcave,and consequentlythemodelproposedbyBindschedleretal. (2010)orbyCailleauetal.(2009a,b)involvingcalcite pseudomorphosisoffungal,bacterial,orplantcell-wall fibrousmaterialseemsmoreappropriateinthatcave.In addition,themodelofBindschedleretal.(2010)couldalso explaintheformationofnanofibers. C ONCLUSIONS Microcalorimetricinvestigationsonmoonmilksamples allowedustogainadditionalinsightsintotheactive Figure6.Induratehyphaeand‘‘filamentshaped’’crystalsproducedby Streptomyces sp.growingonB4agarplates.Notethe similaritiesbetweenthesecrystalsandthefeaturesthatcanbeobservedinmoonmilk.A.Typicalinduratedhyphaewitha smoothappearance.Acentralholemaybeseenonthesestructures(arrow).B.Straightcalcifiedhyphaeshowingindentations duetoepitacticgrowthofcalciumcarbonate.C.Curvedinduratedhyphaeshowingserratededgesduetoepitacticgrowthof calciumcarbonateonlyononeside.D.Epitaxyformingalargecrystalaroundamineralizedhypha;arrowsindicatethehyphae goingthroughthecrystal. M ICROBIOLOGICALACTIVITIESINMOONMILKMONI TOREDUSINGISOTHERMALMICROCALORIMETRY (C AVEOF V ERSCHEZLE B RANDT N EUCHATEL ,S WITZERLAND ) 124 N JournalofCaveandKarstStudies, April2012

PAGE 127

bacterialpopulationasawhole.Thisstudydemonstrates thatitispossibletoestimatethesizeofthemetabolically activemicrobialpopulationinmoonmilksamples.In addition,wedeterminedthegrowthrateofthemetabolicallyactivemicrobialpopulationwithvariouscarbon sourcesadded,aswellasthemaximumconsumptionrate ofthesecarbonsources.Obtaininginformationongrowth ratesandconsumptionratesallowsforinferringwhat processesmightbeimportanttounderstandingthe microbiologyandbiogeochemistryofmoonmilk.Inour study,organicacidswererapidlyconsumed,withcitricacid sustainingthehighestgrowthrate,emphasizingthecrucial roleoforganic-acidinputforthecarbonatemicrofabric processesdrivenbythismicrobialcommunity.Further useofmicrocalorimetry,incombinationwithmolecular techniques,willundoubtedlyleadtoabetterunderstanding ofmoonmilkformationbylinkingmeasurementsofthe rateandextentofmetabolicprocessesandgrowthtothe compositionofthemicrobialcommunity. A CKNOWLEDGEMENTS TheauthorswishtothankDr.PierreVonlanthenatIGP UniLfortechnicalsupportduringSEMobsvervations. Theauthorsarealsothankfulforthevaluablecomments andsuggestionsprovidedbyassociateeditorDr.K.Lavoie andtwoanonymousreviewers. R EFERENCES Amann,R.I.,Ludwig,W.,andSchleifer,H.K.,1995,Phylogenetic identificationandinsitudetectionofindividualmicrobialcells withoutcultivation:MicrobiologicalReviews,v.59,p.143–169. Badis,A.,Ferradji,F.Z.,Boucherit,A.,Fodil,D.,andBoutoumi,H., 2009,Characterizationandbiodegradationofsoilhumicacidsand preliminaryidentificationofdecolorizingactinomycetesatMitidja plainsoils(Algeria):AfricanJournalofMicrobiologyResearch,v.3, p.997–1007. Barros,N.,Feijo o,S.,Simoni,J.A.,Prado,A.G.S.,Barboza,F.D.,and Airoldi,C.,1999,MicrocalorimetricstudyofsomeAmazonian soils:ThermochimicaActa,v.328,p.99–103.doi:10.1016/S00406031(98)00629-7. Barros,N.,Gomez-Orellana,I.,Feijbo o,S.,andBalsa,R.,1995,The effectofsoilmoistureonsoilmicrobialactivitystudiedby microcalorimetry:ThermochimicaActa,v.249,p.161–168.doi: 10.1016/0040-6031(95)90686-X. Barton,H.A.,andNorthup,D.E.,2007,Geomicrobiologyincave environments:past,currentandfutureperspectives:JournalofCave andKarstStudies,v.69,p.163–178. Bindschedler,S.,Millie `re,L.,Cailleau,G.,Job,D.,andVerrecchia,E.P., 2010,Calciticnanofibresinsoilsandcaves:aputativefungal contributiontocarbonatogenesis, in Pedley,H.M.,andRogerson, M.,eds.,TufasandSpeleothems:UnravelingtheMicrobialand PhysicalControls:GeologicalSocietyofLondonSpecialPublications 336,p.225–238.doi:10.1144/SP336.11. Boquet,E.,Boronat,A.,andRamos-Cormenzana,A.,1973,Production ofcalcite(calciumcarbonate)crystalsbysoilbacteriaisageneral phenomenon:Nature,v.246,p.527––529.doi:10.1038/246527a0. Borsato,A.,Frisia,S.,Jones,B.,andVanDerBorg,K.,2000,Calcite moonmilk:crystalmorphologyandenvironmentofformationin cavesintheItalianAlps:JournalofSedimentaryResearch,v.70, p.1171–1182.doi:10.1306/032300701171. Braissant,O.,Cailleau,G.,Aragno,M.,andVerrecchia,E.P.,2004, Biologicallyinducedmineralizationinthetree Miliciaexcelsa (Moraceae):itscausesandconsequencestotheenvironment: Geobiology,v.2,p.59–66.doi:10.1111/j.1472-4677.2004.00019.x. Braissant,O.,Decho,A.W.,Dupraz,C.,Glunk,C.,Przekop,K.M.,and Visscher,P.T.,2007,Exopolymericsubstancesofsulfate-reducing bacteria:InteractionswithcalciumatalkalinepHandimplicationfor formationofcarbonateminerals:Geobiology,v.5,p.401–411.doi: 10.1111/j.1472-4669.2007.00117.x. Braissant,O.,Decho,A.W.,Przekop,K.M.,Gallagher,K.L.,Glunk,C., Dupraz,C.,andVisscher,P.T.,2009,Characteristicsandturnoverof exopolymericsubstancesinahypersalinemicrobialmat:FEMS MicrobiologyEcology,v.67,p.293–307.doi:10.1111/j.1574-6941. 2008.00614.x. Braissant,O.,Verrecchia,E.P.,andAragno,M.,2002,Isthecontribution ofbacteriatoterrestrialcarbonbudgetgreatlyunderestimated?: Naturwissenschaften,v.89,p.366–370.doi:10.1007/s00114-0020340-0. Braissant,O.,Wirz,D.,Go ¨pfert,B.,andDaniels,A.U.,2010,Useof isothermalmicrocalorimetrytomonitormicrobialactivities:FEMS MicrobiologyLetters,v.303,p.1–8.doi:10.1111/j.1574-6968.2009. 01819.x. Bull,A.T.,2004,MicrobialDiversityandBioprospecting,Washington, AmericanSocietyofMicrobiology,496p. Cailleau,G.,Dadras,M.,Abolhassani-Dadras,S.,Braissant,O.,and Verrecchia,E.P.,2009a,Evidenceforanorganicoriginofpedogenic calciticnanofibres:JournalofCrystalGrowth,v.311,p.2490–2495. doi:10.1016/j.jcrysgro.2009.02.029. Cailleau,G.,Verrecchia,E.P.,Braissant,O.,andEmmanuel,L.,2009b, Thebiogenicoriginofneedlefibrecalcite:Sedimentology,v.56, p.1858–1875.doi:10.1111/j.1365-3091.2009.01060.x. Callot,G.,Guyon,A.,andMousain,D.,1985,Inter-relationsentre aiguillesdecalciteethyphesmyce liens:Agronomie,v.5,p.209–216. Can averas,J.C.,Cuezva,S.,Sanchez-Moral,S.,Lario,J.,Laiz,L., Gonzalez,J.M.,andSaiz-Jimenez,C.,2006,Ontheoriginoffiber calcitecrystalsinmoonmilkdeposits:Naturwissenschaften,v.93, p.27–32.doi:10.1007/s00114-005-0052-3. Colwell,R.R.,andGrimes,D.J.,2000,NonculturableMicroorganismsin theEnvironment,Washington,AmericanSocietyofMicrobiology, 354p. Costerton,J.W.,Lewandowski,Z.,Caldwell,D.E.,Korber,D.R.,and Lappin-Scott,H.M.,1995,Microbialbiofilms:AnnualReviewof Microbiology,v.49,p.711–745.doi:10.1146/annurev.mi.49.100195. 003431. Cuezva,S.,Sanchez-Moral,S.,Saiz-Jimenez,C.,andCan averas,J.C., 2009,Microbialcommunitiesandassociatedmineralfabricsin AltamiraCave,Spain:InternationalJournalofSpeleology,v.38, p.83–92. Curry,M.D.,Boston,P.J.,Spilde,M.N.,Baichtal,J.F.,andCampbell, A.R.,2009,Cottonballs,auniquesubaqueousmoonmilkand abundantsubaerialmoonmilkinCataractCave,TongassNational Forest,Alaska:InternationalJournalofSpeleology,v.38, p.111–128. Dari,K.,Be chet,M.,andBlondeau,R.,1995,Isolationofsoil Streptomyces strainscapableofdegradinghumicacidsandanalysis oftheirperoxidaseactivity:FEMSMicrobiologyEcology,v.16, p.115–122.doi:10.1111/j.1574-6941.1995.tb00275.x. Dupraz,C.,Reid,R.P.,Braissant,O.,Decho,A.W.,Norman,R.S.,and Visscher,P.T.,2009,Processesofcarbonateprecipitationinmodern microbialmats:Earth-ScienceReviews,v.96,p.141–162.doi:10. 1016/j.earscirev.2008.10.005. Engel,A.S.,andNorthup,D.E.,2008,CavesandKarstasModelSystems forAdvancingtheMicrobialSciences, in Martin,J.B.,andWhite, W.B.,eds.,FrontiersofKarstResearch,Leesburg,Virginia,Karst WatersInstituteSpecialPublication13,p.37–48. Fujii,K.,Hayakawa,C.,VanHees,P.A.W.,Funakawa,S.,andKosaki, T.,2010,Biodegradationoflowmolecularweightorganiccompounds andtheircontributiontoheterotrophicsoilrespirationinthree Japaneseforestsoils:PlantandSoil,v.334,p.475–489.doi:10.1007/ s11104-010-0398-y. Gaskell,E.E.,Sihanonth,P.,Rostron,C.,Hutcheon,G.A.,andHobbs, G.,2010,Isolationandidentificationofmucinolyticactinomycetes: AntonievanLeeuwenhoek,v.97,p.211–220.doi:10.1007/s10482009-9402-z. O.B RAISSANT ,S.B INDSCHEDLER ,A.U.D ANIELS ,E.P.V ERRECCHIA,AND G.C AILLEAU JournalofCaveandKarstStudies, April2012 N 125

PAGE 128

Gigon,R.,1976,Inventairespe le ologiquedelaSuisse,tome1,Cantonde Neucha ˆtel,Neucha ˆtel,CommissiondeSpe le ologiedelaSocie te helve tiquedesSciencesnaturelles,223p. Goodfellow,M.,andWilliams,S.T.,1983,Ecologyofactinomycetes: AnnualReviewsinMicrobiology,v.37,p.189–216. Higuera-Guisset,J.,Rodr guez-Viejo,J.,Chaco n,M.,Mun oz,F.J., Vigue s,N.,andMas,J.,2005,Calorimetryofmicrobialgrowthusing athermopilebasedmicroreactor:ThermochimicaActa,v.427, p.187–191.doi:10.1016/j.tca.2004.09.010. Holt,J.G.,Krieg,N.R.,Sneath,P.H.A.,Staley,J.T.,andWilliams,S.T., 1994,Bergey’sManualofDeterminativeBacteriology,ninthedition, NewYork,London,LippincottWilliams&Wilkins,791p. James,A.M.,ed.,1987,ThermalandEnergeticStudiesofCellular BiologicalSystems,Bristol,JohnWright,221p. Janssen,P.H.,Yates,P.S.,Grinton,B.E.,Taylor,P.M.,andSait,M., 2002,Improvedculturabilityofsoilbacteriaandisolationinpure cultureofnovelmembersofthedivisionsAcidobacteria,Actinobacteria,Proteobacteria,andVerrucomicrobia:AppliedandEnvironmentalMicrobiology,v.68,p.2391–2396.doi:10.1128/AEM.68.5. 2391-2396.2002. Jones,S.E.,andLennon,J.T.,2010,Dormancycontributestothe maintenanceofmicrobialdiversity:ProceedingsoftheNational AcademyofSciencesoftheUnitedStatesofAmerica,v.107, p.5881–5886.doi:10.1073/pnas.0912765107. Kawai,M.,Yamaguchi,N.,andNasu,M.,1999,Rapidenumeration ofphysiologicallyactivebacteriainpurifiedwaterusedinthepharmaceuticalmanufacturingprocess:JournalofAppliedMicrobiology, v.86,p.496–504.doi:10.1046/j.1365-2672.1999.00689.x. Kimura,T.,andTakahashi,K.,1985,Calorimetricstudiesofsoil microbes:quantitativerelationbetweenheatevolutionduring microbialdegradationofglucoseandchangesinmicrobialactivity insoil:JournalofGeneralMicrobiology,v.131,p.3083–3089.doi: 10.1099/00221287-131-11-3083. Konhauser,K.,2007,IntroductiontoGeomicrobiology,Oxford,BlackwellPublishingCo.,433p. Lacelle,D.,2010,Discussion:‘‘Thebiogenicoriginofneedlefibre calcite’’byG.Cailleauetal.(2009),Sedimentology,56,1858–1875: Sedimentology,v.57,p.1147–1149.doi:10.1111/j.1365-3091.2009. 01134.x. Laiz,L.,Groth,I.,Schumann,P.,Zezza,F.,Felske,A.,Hermosin,B.,and Saiz-Jimenez,C.,2000,MicrobiologyofthestalactitesfromGrotta deiCervi,PortoBadisco,Italy:InternationalMicrobiology,v.3, p.25–30. Muchova ,M.,Ru z ic ka,J.,Julinova ,M.,Dolez alova ,M.,Houser,J., Koutny ,M.,andBun kova ,L.,2009,Xanthanandgellandegradation bybacteriaofactivatedsludge:WaterScienceandTechnology,v.60, p.965–973.doi:10.2166/wst.2009.443. Mulec,J.,2008,Microorganismsinhypogeon:examplesfromSlovenian karstcaves:ActaCarsologica,v.37,p.153–160. Mulec,J.,Zalar,P.,Hajna,N.Z.,andRupnik,M.,2002,Screeningfor culturablemicroorganismsfromcaveenvironments(Slovenia):Acta Carsologica,v.31,no.2,p.177–187. Northup,D.E.,andLavoie,K.H.,2001,Geomicrobiologyofcaves:a review:GeomicrobiologyJournal,v.18,p.199–222.doi:10.1080/ 01490450152467750. Parinkina,O.M.,1973,Determinationofbacterialgrowthratesintundra soils:BulletinsfromtheEcologicalResearchCommittee(Stockholm), v.17,p.303–309. Pas ic ,L.,Kovc e,B.,Sket,B.,andHerzog-Velikonja,B.,2009,Diversityof microbialcommunitiescolonizingthewallsofaKarsticcavein Slovenia:FEMSMicrobiologyEcology,v.71,p.50–60. Perrin,J.,2003,Aconceptualmodelofflowandtransportinakarst aquiferbasedonspatialandtemporalvariationsofnaturaltracers [Ph.D.thesis]:Neuchatel,UniversityofNeuchatel. Phillips,S.E.,Milnes,A.R.,andFoster,R.C.,1987,Calcifiedfilamentsanexampleofbiologicalinfluencesintheformationofcalcrete insouthAustralia:AustralianJournalofSoilResearch,v.25, p.405–428.doi:10.1071/SR9870405. Phillips,S.E.,andSelf,P.G.,1987,Morphology,crystallographyand originofneedlefibercalciteinQuaternarypedogeniccalcretes ofsouthAustralia:AustralianJournalofSoilResearch,v.25, p.429–444.doi:10.1071/SR9870429. Pirt,J.S.,1987,Theenergeticsofmicrobesatslowgrowthrates: maintenanceenergiesanddormantorganisms:JournalofFermentationTechnology,v.65,p.173–177.doi:10.1016/0385-6380(87) 90161-0. Portillo,M.C.,andGonzalez,J.M.,2011,Moonmilkdepositsoriginate fromspecificbacterialcommunitiesinAltamiraCave(Spain): MicrobialEcology,v.61,p.182–189.doi:10.1007/s00248-010-9731-5. Portillo,M.C.,Gonzalez,J.M.,andSaiz-Jimenez,C.,2008,Metabolically activemicrobialcommunitiesofyellowandgreycolonizationsonthe wallsofAltamiraCave,Spain:JournalofAppliedMicrobiology, v.104,p.681–691.doi:10.1111/j.1365-2672.2007.03594.x. Qian,F.,An,L.,Wang,M.,Li,C.,andLi,X.,2007,Isolationand characterizationofaxanthan-degrading Microbacterium sp.strainXT11 fromgardensoil:JournalofAppliedMicrobiology,v.102,p.1362–1371. Rohwerder,T.,Sand,W.,andLascu,C.,2003,Preliminaryevidencefora sulphurcycleinMovileCave,Romania:ActaBiotechnologica,v.23, p.101–107.doi:10.1002/abio.200390000. Rong,X.-M.,Huang,Q.-Y.,Jiang,D.-H.,Cai,P.,andLiang,W.,2007, Isothermalmicrocalorimetry:areviewofapplicationsinsoiland environmentalsciences:Pedosphere,v.17,p.137–145.doi:10.1016/ S1002-0160(07)60019-8. Sahin,N.,2003,Oxalotrophicbacteria:ResearchinMicrobiology,v.154, p.399–407.doi:10.1016/S0923-2508(03)00112-8. Sahin,N.,2004,Isolationandcharacterizationofmesophilic,oxalatedegrading Streptomyces fromplantrhizosphereandforestsoils: Naturwissenschaften,v.91,p.498–502.doi:10.1007/s00114-004-0562-4. Segawa,T.,Yoshimura,Y.,Watanabe,K.,Kanda,H.,andKohshima,S., 2011,Communitystructureofculturablebacteriaonsurfaceof GulkanaGlacier,Alaska:PolarScience,v.5,p.41–51.doi:10.1016/ j.polar.2010.12.002. Shaw,W.V.,andHopwood,D.A.,1976,Chloramphenicolacetylationin Streptomyces :JournalofGeneralMicrobiology,v.94,p.159–166. doi:10.1099/00221287-94-1-159. Stomeo,F.,Portillo,M.C.,Gonzalez,J.M.,Laiz,L.,andSaiz-Jimenez, C.,2008, Pseudonocardia inwhitecolonizationsintwocaveswith Paleolithicpaintings:InternationalBiodeterioration&Biodegradation,v.62,p.483–486.doi:10.1016/j.ibiod.2007.12.011. Tan,H.,Deng,Z.,andCao,L.,2009,Isolationandcharacterizationof actinomycetesfromhealthygoatfaeces:LettersinAppliedMicrobiology,v.49,p.248–253.doi:10.1111/j.1472-765X.2009.02649.x. vanHees,P.A.W.,Jones,D.L.,andGodbold,D.L.,2002,Biodegradation oflowmolecularweightorganicacidsinconiferousforestpodzolic soils:SoilBiologyandBiochemistry,v.34,p.1261–1272.doi: 10.1016/S0038-0717(02)00068-8. Wadso ¨,I.,2002,Isothermalmicrocalorimetryinappliedbiology:ThermochimicaActa,v.394,p.305–311.doi:10.1016/S0040-6031(02)00263-0. Wadso ¨,I.,2009,Characterizationofmicrobialactivityinsoilbyuse ofisothermalmicrocalorimetry:JournalofThermalAnalysisand Calorimetry,v.95,p.843–850.doi:10.1007/s10973-008-9467-3. M ICROBIOLOGICALACTIVITIESINMOONMILKMONI TOREDUSINGISOTHERMALMICROCALORIMETRY (C AVEOF V ERSCHEZLE B RANDT N EUCHATEL ,S WITZERLAND ) 126 N JournalofCaveandKarstStudies, April2012

PAGE 129

IMPORTANCEOFKARSTSINKHOLESINPRESERVING RELICT,MOUNTAIN,ANDWET-WOODLANDPLANT SPECIESUNDERSUB-MEDITERRANEANCLIMATE: ACASESTUDYFROMSOUTHERNHUNGARY Z OLTA N B A TORI 1 *,L A SZLO K O ¨ RMO ¨ CZI 1 ,L A SZLO E RDO S 1 ,M A RTA Z ALATNAI 1 AND J A NOS C SIKY 2 Abstract: Speciescompositionandthevegetationpatternoftheunderstorywere investigatedindifferentsizedsolutionsinkholesinawoodlandareaoftheMecsek Mountains(southernHungary).Vegetationdatatogetherwithtopographicvariables werecollectedalongtransectstorevealthevegetationpatternsontheslopes,anda specieslistwascompiledforeachsinkhole.Theresultsindicatethatthevegetation patternsignificantlycorrelateswithsinkholesize.Insmallersinkholes,vegetationdoes notchangesubstantiallyalongthetransects;inlargersinkholes,however,vegetation inversionispronounced.Wealsofoundthatsinkholesizeclearlyinfluencesthenumber ofvascularplantspecies,inaccordancewiththewell-knownrelationshipbetweenspecies numberandarea.Intheforestlandscape,manymedium-sizedandlargesinkholeshave developedintoexcellentrefugeareasforglacialrelicts,mountain,andwet-woodland plantspecies. I NTRODUCTION Climate-inducedspeciesextinctionhasbecomeamajor topicinconservationbiology.Articlesandbooksfocusingon climatechangehaveappeared(e.g.,IversonandPrasad,1998; Sagarinetal.,1999;Cowie,2007),andarapidlyincreasing amountofinformationisavailableaboutcurrentand potentialrefugeareas(e.g.,Ko ¨hnandWaterstraat,1990; Schindleretal.,1996;Sheldonetal.,2008)wheremany speciesmaysurviveunfavorableregionalenvironmental conditions.Duringglacialperiods,amajorpartofEurope waslargelycoveredbycoldhabitats,andonlycold-adapted specieswereabletosurviveundertheseextremeconditions (Habeletal.,2010).However,afterglacialretreat,siteswith coldandhumidclimatesbecameimportanttopreserving glacialrelictsandhighmountainandmountainspecies, mainlyinlowermountainandhillranges. Onaglobalscale,extensivekarstlimestonebedrock playsanimportantroleinthepreservationofrare, endangered,orspecializedspecies(e.g.,Christiansenand Bellinger,1996;Wolowski,2003;Judson,2007;Lewisand Bowman,2010).Karstlandformslikecaves,wellsand sinkholes(alsoknownasdolines)determinethegeomorphologic,microclimatic,andvegetationfeaturesofkarst surfacesandinfluencethekarstaquifersystem.Moreover, cavesandwellsarehotspotsofsubterraneanbiodiversity (CulverandSket,2000;Elliott,2007);sinkholespreserverelicts(Horvat,1953;Lazarevic etal.,2009),high mountain,mountain,(Beckv.Mannagetta,1906;Horvat, 1953;PericinandHu ¨rlimann,2001;Dakskobleretal., 2008)andendemic(Eglietal.,1990;BrulloandGiussodel Galdo,2001;O ¨ zkanetal.,2010)species,and,inmany cases,theyareanimportantsourceofknowledgeabout vegetationhistory.Forexample, Dracocephalumruyschiana ,aglacialrelictinthesinkholefloraofnorthern Hungary,indicatesaformerperiglacialclimate(Kira ly, 2009),butsomehighmountainelements(e.g. Lilium martagon subsp. alpinum Ribesalpinum )alsooccurinthe low-lyingsinkholes(between400and600masl)ofthearea (Szmorad,1999;Vojtko ,1997). Understandingthepatternsofsinkholevegetation requiresanunderstandingofthesurroundingvegetation patterns.AccordingtoHorvat(1953),thecoolandhumidmicroclimateofsinkholesmayaffecttheirfloraand vegetationintwodifferentways.Inmanycases,thermal inversionleadstoaninversionofsurroundingvegetation zones.Ontheotherhand,edaphicvegetationtypesmay alsoappearonthebottomofsinkholesunderspecial ecologicalconditions(Egli,1991;Ba torietal.,2009).From anecologicalpointofview,thelatterismoreimportant,as itmayprovideprimaryhabitatsformanyspeciesabsentin thesurroundingvegetation. Thepurposeofthepresentstudyistodetermineand comparethevegetationpatternandspeciescomposition insolutionsinkholesofthesub-Mediterraneanpartof Hungarywithregardtosinkholesizeandtooffersome usefulexplanationsfortheirroleinnatureconservation. Thefollowingquestionsareaddressed:(i)Whatisthe extentofvegetationinversionindifferent-sizedsinkholesin awoodlandarea?(ii)Howdoestheextentofrefugeareas changewithsinkholesize?(iii)Howmanyrelict,mountain, *Correspondingauthor:zbatory@gmail.com 1 DepartmentofEcology,UniversityofSzeged,6726Szeged,Ko ¨ze pfasor52, Hungary 2 DepartmentofPlantTaxonomyandGeobotany,UniversityofPe cs,7624Pe cs, Ifju sa gu tja6,Hungary Z.Ba tori,L.Ko ¨rmo ¨czi,L.Erdo s,M.Zalatnai,andJ.Csiky–Importanceofkarstsinkholesinpreservingrelict,mountain,andwetwoodlandplantspeciesundersub-Mediterraneanclimate:AcasestudyfromsouthernHungary. JournalofCaveandKarstStudies, v.74, no.1,p.127–134.DOI:10.4311/2011LSC0216 JournalofCaveandKarstStudies, April2012 N 127

PAGE 130

andwet-woodlandplantspeciescanbefoundinthe differentsizedsinkholes? M ETHODS Thestudywascarriedoutinthekarstareaof30km 2 in theMecsekMountains(southernHungary),nearthecity ofPe cs(Fig.1).Onthekarstsurface,therearemorethan twothousandsinkholeslocatedbetween250and500m abovesealevel(Fig.2).Theformationofthesedepressions startedduringthePleistocene,anditisstillintensivedue totheabundantprecipitationonthefissuredbedrock underlyingthewoodland.Thediameterofthelargest sinkholeisover200manditsdepthexceeds30m(Lova sz, 1971),butmorethanfifteenhundredofthesesinkholesare quitesmall(diameter 20m).Thesinkholedensityofthis areaisextremelyhigh,withthemaximumof380sinkholes perkm 2 .Thesedepressionsareinprimitivestagesof development,shownbytheirsteepslopesandafunnel-like form(Hoyk,1999). Theaverageannualrainfallofthestudysiteexceeds 700mm,withconsiderableinterannualvariation.Dueto thesub-Mediterraneanclimate,themonthlymaximum valuesoccurduringsummerandautumn(MayandJune 77mm,October72mm).Theannualmeantemperatureis about8.8 u C,withthehighestmonthlymeantemperature of19.3 u CinJuly.Wintersaremoderatelycoldwith 2 1.1 u C meantemperaturefromDecembertoFebruary(A da m etal.,1981). Sub-Mediterraneantype,middle-aged(70to110years old)mixed-oakandbeechforestsdominatethepresent vegetationoftheplateausandslopesofthestudysite.The mostimportantAtlantic-Mediterranean,sub-Mediterranean, andMediterraneanplantsinclude Aremoniaagrimonoides Asperulataurina Helleborusodorus Lathyrusvenetus Luzula forsteri Potentillamicrantha Rosaarvensis Ruscusaculeatus Ruscushypoglossum Scutellariaaltissima Tamuscommunis and Tiliatomentosa .Hotspotsofmountainspeciesofthearea canbefoundintheforestsofthedeep,humid,androcky ravinesandvalleys. Oursurveyswereconductedbetween2006and2011 fromearlyJunetomid-Septemberonthekarstsurfaceof theMecsekMountains.Sinkholeswereselectedinsites thatdidnotshowsignsofrecentwood-cutting.Sinkholes rankedbydiameterareidentifiedwithcapitallettersfrom AtoT(Table1). Transectsforsamplingunderstorywereestablished acrossthetwentysinkholesinanorth-southdirection, passingthroughthedeepestpointofthedepressions. Transectsconsistedofseriesof1msquarecontiguous plots.Inthelargersinkholes,thetransectswere2mwide, withindividualplotsside-by-sidealongthem.Inthe smallersinkholes,onlya1mwideserieswassurveyed. Percentagecoverofeachvascularplantspecieswas estimatedvisuallyintheplots.Furthermore,afloralist foreachsinkholewascompletedbyasystematicsearch throughthetotalareaofthesinkhole.Thetotalarea includedtheareaoftheslopes(wheretheslopeanglewas over10 u )andtheareaoftheedges,anapproximately1to 5mwidestriparoundthesmallersinkholes,andan approximately10to20mwidestriparoundthelarger sinkholeswheretheslopeanglewaslessthan10 u .For comparison,4051msquareplotswererandomlytaken fromthethreehabitattypes,mixed-oakforests,beech Figure1.LocationofthestudysiteintheMecsek Mountains(southernHungary). Figure2.Alargesolutionsinkhole(sinkholeO)ofthe MecsekMountainsinthewinterof2008. I MPORTANCEOFKARSTSINKHOLESINPRESERVINGRELICT,MOUNTAIN,ANDWET-WOODLANDPLANTSPECIESUNDERSUB -M EDITERRANEANCLIMATE: A CASESTUDYFROMSOUTHERN H UNGARY 128 N JournalofCaveandKarstStudies, April2012

PAGE 131

forests,andforestsinthedeepravinesandvalleys(called ravineforestsbelow),occurringintheneighborhoodofthe sinkholes.Westudiedatotalof4017plotsonthekarstof MecsekMountains:3612plotsinsinkholesand405plots intheirsurroundings.Atotalof251vascularplantspecies wereincludedintheanalyses.Inaddition,thediameterand depthofthesinkholesweremeasured(Table1).Plant specieswerenamedaccordingtoSimon(2000). Diagnostic(ordifferential)speciesincludespecieswith highoccurrencewithinagivenvegetationtype(orwithin somevegetationtypes)andlowoccurrenceinothervegetationtypes.Diagnosticspeciesofthedifferentforesttypes surroundingthesinkholesweredeterminedbystatistical fidelitymeasures(Chytry etal.,2002).Thephicoefficient( W ) forallspecieswascomputedwiththeJUICE7.0.25program (Tichy ,2002).Thiscoefficientrangesfrom 2 1to1,butfor convenience,itismultipliedby100intheprogram.The highestphivalueof1isachievedifthespeciesoccursinall plotsofthetargetvegetationtypeandisabsentelsewhere. Forthecomparisonofmixed-oakforests,beechforests,and ravineforests(405totalplots),weusedthe W $ 0.1threshold (Fisher’sexacttest, p 0.05)duringanalyses.Theclassified datasetcontainedonlysitegroupsofequalsize(135plotsfor eachgroup).Inthisscale,fidelitymeasurementresultedin threegroupsofdiagnosticspeciesintheunderstoriesofthe mixed-oakforests,thebeechforests,andtheravineforests (Table2).Finally,weclassifiedeach1msquareplotalong thesinkholetransectsintoplottypeswiththeuseofthe diagnosticspeciesgroups.Forexample,ifthenumberof diagnosticspeciesoftheravineforestswasthehighestinthe targetsinkholeplot,weconsidereditaplotdominatedby ravineforestspecies.Hence,thismethodresultedinfive sinkhole-plottypes:plotsdominatedbyravineforestspecies, plotsdominatedbybeechforestspecies,plotsdominatedby mixed-oakforestspecies,transitionalplots,andemptyplots. Transitionalplotswerethosewhichweredominatedbyan equalnumberofdiagnosticspeciesoftwoorthreegroups. Emptyplotsdidnotcontainanyplantspecies.Froman ecologicalpointofview,plotsdominatedbyravineforest speciescanbeconsideredpotentialrefugeareasformany speciesadaptedtocoolandmoisthabitats(e.g.,Fig.3). Moreover,fieldobservationswereperformedtoverifythe resultoffidelitymeasurement. Species-arearelationswereassessedforallplantspecies, aswellasforthegroupofrelict,mountain,wet-woodland speciesandotherdiagnosticspeciesoftheravineforests.In thisstudy,twentysinkholes,representativeofsinkholesof allsizes,wereincludedinthespecies-areaanalyses. R ESULTS Thenumberof1msquareplotsdominatedbyravine forestspeciesgenerallyincreasedwithsinkholesize(Fig.4; Table3).InthesmallandshallowsinkholesAtoG,only plotsdominatedbybeechforestandmixed-oakforest specieswerefound.Incontrast,inlargersinkholes,the lowerpartsoftheslopeswerecoveredmainlybyplots dominatedbyravineforestspecies;whiletheupperparts oftheslopeswerecoveredbyplotsdominatedbybeech forestormixed-oakforestspecies.Insomesinkholes,the proportionofmixed-oakforestspecieswasveryhighon theedgesandsouth-facingslopes.Accordingly,vegetation inversionwaspronounced,especiallyinthecaseof sinkholesN,S,Q,andT,followedbyagradualchange infloristiccomposition. Oftheforty-onediagnosticspeciesoftheravineforests found,fiveweremountainspecies,tenwerewet-woodland species,andtwenty-sixwereotherdiagnosticspecies(e.g., gap-species)(Table2).Inaddition,aglacialrelict( Stachys alpina ),sixothermountainspecies( Actaeaspicata Aruncus sylvestris Dryopterisaffinis Dryopterisdilatata Dryopteris expansa ,and Polystichum 3 bicknelli ),andsixotherwetwoodlandspecies( Deschampsiacaespitosa Eupatorium cannabinum Festucagigantea Lycopuseuropaeus Rumex sanguineus ,and Solanumdulcamara )werefoundinthe ravineforests,althoughtheywerenotdiagnosticaccording tothetest. Therelationshipbetweensinkholesize(log 10 transformed)andspeciesnumber(log 10 transformed)isshown inFig.5.Whenallspecieswereconsidered,thecorrelationbetweenspeciesnumberandsinkholesizewas positiveandsignificant( R 2 5 0.9302, p 0.001).For example,thehighestnumberofspecies(141)wasfound inthelargestsinkhole(T),whilethelowestnumberof species(23)wasfoundintwoofthesmallestsinkholes(A andC)(Table3).Theresultwasbasicallythesameifonly thegroupofrelict,mountain,wet-woodlandspecies andotherdiagnosticspeciesoftheravineforestswas considered( R 2 5 0.9006, p 0.001).Fromafloristic pointofview,sinkholesRandTwerethemostimportant, becausetheycontainedthehighestnumberofbothrelicts andmountainspecies(R: Dryopterisaffinis Dryopteris expansa Dryopterisdilatata Polystichumaculeatum Polystichum 3 bicknelli and Stachysalpina ;T: Aconitum vulparia Actaeaspicata Aspleniumscolopendrium Dryopterisaffinis Polystichumaculeatum Polystichum 3 bicknelli and Stachysalpina ). Table1.DiametersanddepthsofthestudiedsinkholesoftheMecsekMountains. Sinkhole Dimensions,mABCDEFGHIJKLMNOPQRST Diameter9.514.51518212343606976818592124135145158167187229 Depth12.52.534.53.5712121215151317192522222131 Z.B A TORI ,L.K O ¨ RMO ¨ CZI ,L.E RDO S ,M.Z ALATNAI AND J.C SIKY JournalofCaveandKarstStudies, April2012 N 129

PAGE 132

Table2.Diagnosticspeciesofthetypesofforestsurrounding sinkholesintheMecsekMountains,definedby W 3 100 $ 10.Ifaspeciesisinthelistformorethanoneforesttype,its W valueisshowninbothpartsofthetable. VegetationTypeandSpecies W 3 100 MOFBFRF Diagnosticspeciesofthemixed-oakforests Lathyrusvernus 10.321.5 ??? Festucadrymeja 11.1 ?????? Melittiscarpatica 11.1 ?????? Galiummollugo 12.2 ?????? Poanemoralis 12.2 ?????? Moehringiatrinervia 12.9 ?????? Campanulapersicifolia 14.1 ?????? Carexdivulsa 14.1 ?????? Symphytumtuberosum 14.1 ?????? Torilisjaponica 14.1 ?????? Brachypodiumsylvaticum 14.2 ?????? Sorbustorminalis 14.2 ?????? Galiumaparine 15.8 ?????? Luzulaforsteri 15.8 ?????? Prunellavulgaris 15.8 ?????? Alliariapetiolata 16.2 ?????? Ruscusaculeatus 16.3 ?????? Carexpilosa 16.451.1 ??? Quercuspetraea 18.414.2 ??? Potentillamicrantha 20.1 ?????? Violaalba 20.1 ?????? Violaodorata 20.1 ?????? Buglossoidespurpureocoerulea 22.5 ?????? Lysimachianummularia 22.5 ?????? Veronicachamaedrys 23.3 ?????? Tamuscommunis 23.6 ?????? Festucaheterophylla 24.7 ?????? Fragariavesca 26.4 ?????? Galiumschultesii 26.8 ?????? Geumurbanum 29.1 ?????? Campanularapunculoides 30.5 ?????? Crataeguslaevigata 30.6 ?????? Acercampestre 30.8 ?????? Carpinusbetulus 31.3 ?????? Helleborusodorus 33.3 ?????? Quercuscerris 34.4 ?????? Convallariamajalis 34.7 ?????? Euphorbiaamygdaloides 35.0 ?????? Fallopiadumetorum 39.3 ?????? Euonymusverrucosus 40.0 ?????? Bromusramosus agg.43.5 ?????? Melicauniflora 45.139.8 ??? Rosaarvensis 45.3 ?????? Glechomahirsuta 54.9 ?????? Ligustrumvulgare 56.9 ?????? Stellariaholostea 60.6 ?????? VegetationTypeandSpecies W 3 100 MOFBFRF Fraxinusornus 66.4 ?????? Dactylispolygama 72.9 ?????? Diagnosticspeciesofthebeechforests Ulmusglabra ??? 12.4 ??? Miliumeffusum ??? 14.1 ??? Quercuspetraea 18.414.2 ??? Galeobdolonluteum s.l. ??? 14.759.8 Asarumeuropaeum ??? 15.1 ??? Violareichenbachiana ??? 15.1 ??? Carexdigitata ??? 16.3 ??? Hepaticanobilis ??? 20.9 ??? Lathyrusvernus 10.321.5 ??? Ruscushypoglossu ??? 21.8 ??? Prunusavium ??? 24.7 ??? Tiliatomentosa ??? 25.6 ??? Tiliacordata ??? 31.4 ??? Rubushirtus agg. ??? 32.4 ??? Fraxinusexcelsior ??? 32.6 ??? Melicauniflora 45.139.8 ??? Hederahelix ??? 43.5 ??? Fagussylvatica ??? 49.2 ??? Carexpilosa 16.151.1 ??? Galiumodoratum ??? 53.9 ??? Diagnosticspeciesoftheravineforests Mountainspecies Polystichumaculeatum ?????? 22.5 Aconitumvulparia ?????? 26.8 Lunariarediviva ?????? 27.7 Silenedioica ?????? 30.5 Aspleniumscolopendrium ?????? 37.8 Wetwoodlandspecies Dryopteriscarthusiana ?????? 15.8 Myosotonaquaticum ?????? 15.8 Persicariadubia ?????? 15.8 Carexpendula ?????? 17.3 Ranunculusrepens ?????? 25.8 Carexremota ?????? 27.7 Urticadioica ?????? 42.8 Aegopodiumpodagraria ?????? 44.8 Athyriumfilix-femina ?????? 46.2 Chrysosplenium alternifolium ?????? 64.9 Otherdiagnosticspeciesoftheravineforests Mercurialisperennis ?????? 11.1 Galeopsisspeciosa ?????? 12.2 Erigeronannuus ?????? 12.2 Pulmonariaofficinalis ?????? 12.4 Cerastiumsylvaticum ?????? 14.1 Knautiadrymeia ?????? 14.1 Table2.Continued. I MPORTANCEOFKARSTSINKHOLESINPRESERVINGRELICT,MOUNTAIN,ANDWET-WOODLANDPLANTSPECIESUNDERSUB -M EDITERRANEANCLIMATE: A CASESTUDYFROMSOUTHERN H UNGARY 130 N JournalofCaveandKarstStudies, April2012

PAGE 133

D ISCUSSIONAND C ONCLUSIONS Studyingtheecologicalconditions,vegetationpattern, andspeciescompositionofkarstdepressionsmayprovide importantinformationfortheconservationandmanagementofkarstsurfaces.Manystudiesinvegetationscience havefocusedonthelarge-scalevegetationpatternsofkarst forms(e.g.Lausi,1964;Ba torietal.,2009),butonly relativelyfewstudiesdealwiththefine-scalevegetation patternofsinkholesandtheuseoftransectstorevealthem (e.g.,Garganoetal.,2010).Inourstudy,thespecies compositionandvegetationpatternofsolutionsinkholes wereanalyzedinrelationtosinkholesizeinasubMediterraneanwoodlandareaofHungary.Thevegetation patternofsinkholescanbesummarizedandthequestions posedintheIntroductioncanbeansweredasfollows. (i)Vegetationinversionisawellknownphenomenonin karstdepressions(Beckv.Mannagetta,1906;Lausi, 1964;Horvat,1953;FavrettoandPoldini,1985).In Hungary,south-facingslopesreceivemuchmoresolar radiation,andthus,arewarmerthannorth-facing slopes(Jakucs,1977).Insmallersinkholesofthestudy site,thevegetationpatterndoesnotchangesubstantiallyalongthetransects,andtheslopesarefloored byvegetationcharacteristiconlyofbeechforestsor mixed-oakforests.Incontrast,south-facingslopesin largersinkholesaredominatedbymixed-oakforests orbeechforests,north-facingslopesbybeechforests, andthebottomofsinkholesbyravineforests. Vegetationinversioniswellpronouncedonlyinthe largestsinkholes,wherebeechforestvegetation replacesthatofmixed-oakforestsonthedeeperparts oftheslopes.Thisphenomenonwasalsoconfirmed byfieldobservations.Similarresultswerepublished byBa torietal.(2009,2011)indifferentscales. (ii)Climatechangehasalreadyproducedandwill continuetoproducenumerousshiftsinthedistributionsofspecies(Waltheretal.,2002),whichhighlights theroleofcurrentandpotentialrefugeareas.A prominentfindinginourstudyisthatthelow-lying sinkholes(250to500masl)oftheMecsekMountains providegoodrefugeareasformanyspeciesadaptedto coolandmoisthabitats.Thisisaconsequenceof themorphologiccharacteristicsofkarstdepressions, whichstronglydeterminebothabiotic(e.g.,air humidity,airtemperature,soilmoisture)andbiotic (e.g.,vegetationpattern)parametersofsinkholes(see also,Geiger,1950;Antonic etal.,1997;Ba ra ny-Kevei, 1999;Antonic etal.,2001,Whitemanetal.,2004; Garganoetal.,2010).Theextentofrefugeareas showsapositivecorrelationwithsinkholesizeinthe MecsekMountains.Ingeneral,theextentofcooland moisthabitatsinthesinkholesincreaseswithsinkhole diameter,duetothefactthatwidersinkholesare usuallydeeper. (iii)SinkholesoftheMecsekMountainsharbormany vascularplantspeciesthataremissingorarevery rareinthesurroundinghabitats,andtheycanbe consideredhabitatislandsinthe‘‘ocean’’oflocal VegetationTypeandSpecies W 3 100 MOFBFRF Atropabella-donna ?????? 15.8 Salviaglutinosa ?????? 15.8 Clematisvitalba ?????? 16.3 Lamiummaculatum ?????? 16.6 Chelidoniummajus ?????? 18.8 Geraniumphaeum ?????? 18.8 Stachysylvatica ?????? 20.1 Cardamineimpatiens ?????? 20.4 Polystichumsetiferum ?????? 21.3 Aspleniumtrichomanes ?????? 26.8 Veronicamontana ?????? 27.4 Geraniumrobertianum ?????? 27.5 Sambucusnigra ?????? 31.4 Dryopterisfilix-mas ?????? 32.2 Mycelismuralis ?????? 33.3 Stellariamedia ?????? 33.9 Acerpseudoplatanus ?????? 35.5 Circaealutetiana ?????? 52.2 Oxalisacetosella ?????? 54.7 Galeobdolonluteum s.l. ??? 14.759.8 Totals No.ofplots135135135 No.ofdiagnosticspecies482041 NoteAbbreviations:MOF:mixed-oakforests,BF:beechforests,RF:ravineforests. Figure3. Stachysalpina (left),aglacialrelict,and Lunaria rediviva (right),amountainspecies,inthesinkholesofthe MecsekMountains,southernHungary. Table2.Continued. Z.B A TORI ,L.K O ¨ RMO ¨ CZI ,L.E RDO S ,M.Z ALATNAI AND J.C SIKY JournalofCaveandKarstStudies, April2012 N 131

PAGE 134

Table3.Totalnumberofplantspeciesandcool-adaptedspeciesofvarioustypes,aswellasthenumberofplotsdominatedby ravineforestspeciesforthesinkholes,whichincreaseinsizefromAtoT. SinkholeABCDEFGHIJKLMNOPQRST Totalnumber ofplant species233723303243395868686672639782918693113141 Glacialrelicts ????????????????????????????????????????????? 1 ??? 1 ??? 1 Mountain species ????????????????????? 1212223233536 Wetwoodland species ????????????????????? 2342327566868 Other diagnostic speciesofthe ravineforests 432248711121110111216151316161618 Plots dominatedby ravineforest species ????????????????????? 822202546253910979596165107 Figure4.Plotsinthenorth-southtransects,withnorthtotheleft,thataredominatedbymixed-oakforest(red),beechforest (green),orravineforest(blue)speciesinthesinkholes(A-T)oftheMecsekMountains.Whiteplotsaretransitionalorempty. Shortverticallinesindicatewheretheslopefallsbelow10 6 attheedgesofthesinkholes,andarrowsmarkthedeepestpointof thesinkholes,whereslopeexposurechanges. I MPORTANCEOFKARSTSINKHOLESINPRESERVINGRELICT,MOUNTAIN,ANDWET-WOODLANDPLANTSPECIESUNDERSUB -M EDITERRANEANCLIMATE: A CASESTUDYFROMSOUTHERN H UNGARY 132 N JournalofCaveandKarstStudies, April2012

PAGE 135

beechandmixed-oakforests.Accordingtothewellknownspecies-arearelationship(Arrhenius,1921), speciesnumberisrelatedtoareabythefunction S 5 CA z ,where S isspeciesnumber, A isareaofisland, and C and z arepositiveconstants. C and z constants werecalculatedbylinearregressiononthelogarithmic formoftheequation,log S 5 log C + z log A ,where log C representsthe y -interceptand z theslope.When allspeciesofthestudiedsinkholesareconsidered,the z valueis0.26,whichisingoodagreementwiththe z valuesreceivedformanyoceanicandhabitatislands ( z 5 0.20to0.35)inislandbiogeography(MacArthur andWilson,1967;SimberloffandAbele,1976;Begon etal.,2005).Incontrast,whenonlythegroupofrelict, mountain,wet-woodlandspeciesandotherdiagnostic speciesoftheravineforestsisconsidered,the z value isconsiderablyhigher( z 5 0.45).Accordingto Rockwood(2006), z valueslargerthanexpectedarise whenislandshavealargehabitatdiversityandare moreorlessisolated.Forexample,Culveretal.(1973) foundarelativelyhigh z valueforterrestrial invertebratesincaves( z 5 0.72),Trejo-Torresand Ackerman(2001)forendemicorchidspecieson geologicallydiversemontaneislands( z 5 0.68),and Brown(1971)forsmallborealmammalsonisolated mountaintops( z 5 0.43).Accordingly,ourresults suggestthatthehabitattopographyoflargesinkholes iscomplexandtheextentofcoolandmoisthabitats considerablyincreaseswithsinkholesize(Fig.4),so largersinkholesmaypreservemanymorevascular plantspeciesadaptedtocoolandmoisthabitatsthan smallersinkholes. Therefore,conservationmanagementmustfocuson protectinghabitatsoflargersinkholesandtheirsurroundingsintheMecsekMountains.Thismanagementshould includeestablishingabufferzonearoundallsinkholes, inaccordancewiththeproposaloftheForestSinkhole Manual(Kiernan,2002). A CKNOWLEDGMENTS WewouldliketothankAndra sVojtko ,Sa ndorBartha andTama sMorschhauserforusefulcommentsandsuggestions.ThisresearchwassupportedbytheTA MOP4.2.2/08/1/2008-0008andtheTA MOP-4.2.1/B-09/1/KONV2010-0005programsoftheHung arianNationalDevelopment Agency. R EFERENCES A da m,L.,Marosi,S.,andSzila rd,J.,eds.,1981,ADuna ntu li-dombsa g (De l-Duna ntu l).Magyarorsza gta jfo ¨ldrajza4:Budapest,Akade miai Kiado ,704p. Antonic ,O.,Kus an,V.,andHras ovec,B.,1997,Microclimaticand topoclimaticdifferencesbetweenthephytocoenosesintheViljska PonikvaSinkhole,Mt.Risnjak,Croatia:HrvatskiMeteorolos ki c asopis,v.32,p.37–49. Antonic ,O.,Hatic,D.,andPernar,R.,2001,DEM-baseddepthinsinkas anenvironmentalestimator:EcologicalModeling,v.138,p.247–254. doi:10.1016/S0304-3800(00)00405-1. Arrhenius,O.,1921,Speciesandarea:JournalofEcology,v.9,p.95–99. Ba ra ny-Kevei,I.,1999,Microclimateofkarsticdolines:ActaClimatologicaUniversitatisSzegediensis,v.32–33,p.19–27. Ba tori,Z.,Galle ,R.,Erdo s,L.,andKo ¨rmo ¨czi,L.,2011,Ecological conditions,floraandvegetationofalargedolineintheMecsek Mountains(SouthHungary):ActaBotanicaCroatica,v.70, p.147–155.doi:10.2478/v10184-010-0018-1. Ba tori,Z.,Csiky,J.,Erdo s,L.,Morschhauser,T.,To ¨ro ¨k,P.,and Ko ¨rmo ¨czi,L.,2009,VegetationofthedolinesinMecsekMountains (SouthHungary)inrelationtothelocalplantcommunities:Acta Carsologica,v.38,no.2–3,p.237–252. Beckv.Mannagetta,G.,1906,DieUmkehrungderPflanzenregionenin denDolinendesKarstes:SitzungsberichtederKaiserlicheAkademie derWissenschafteninWien–Mathematisch-Naturwissenschaftliche Klasse,v.115,p.3–20. Begon,M.,Townsend,C.R.,andHarper,J.L.,2006,Ecology:From IndividualstoEcosystems:Oxford,Blackwell,738p. Brown,J.H.,1971,Mammalsonmountaintops:nonequilibriuminsular biogeography:TheAmericanNaturalist,v.105,no.945,p.467–478. Brullo,S.,andGiussodelGaldo,G.,2001, Astracanthadolinicola (Fabaceae),anewspeciesfromCrete:NordicJournalofBotany, v.21,p.475–480.doi:10.1111/j.1756-1051.2001.tb00799.x. Christiansen,K.,andBellinger,P.,1996,Cave Pseudosinella and Oncopodura newtoscience:JournalofCaveandKarstStudies, v.58,no.1,p.38–53. Chytry ,M.,Tichy ,L.,Holt,J.,andBotta-Duka t,Z.,2002,Determination ofdiagnosticspecieswithstatisticalfidelitymeasures:Journalof VegetationScience,v.13,no.1,p.79–90.doi:10.1111/j.1654-1103. 2002.tb02025.x. Cowie,J.,2007,ClimateChange:BiologicalandHumanAspects:New York,CambridgeUniversityPress,504p. Culver,D.C.,Holsinger,J.R.,andBaroody,R.,1973,Towardapredictive cavebiogeography:theGreenbriervalleyasacasestudy:Evolution, v.27,p.689–695. Figure5.Relationshipbetweensinkholearea(log 10 transformed)andspeciesnumber(log 10 transformed)forvascular vegetationoftheMecsekMountains( N = 20).Species-area linesweredeterminedforallspecies(A: y = 0.2612 x + 0.8670, R 2 = 0.9302)andthegroupofrelict,mountain,wetwoodlandspeciesandotherdiagnosticspeciesoftheravine forests(B: y = 0.4465 x 2 0.4868, R 2 = 0.9006). Z.B A TORI ,L.K O ¨ RMO ¨ CZI ,L.E RDO S ,M.Z ALATNAI AND J.C SIKY JournalofCaveandKarstStudies, April2012 N 133

PAGE 136

Culver,D.C.,andSket,B.,2000,Hotspotsofsubterraneanbiodiversityin cavesandwells:JournalofCaveandKarstStudies,v.62,no.1, p.11–17. Dakskobler,I.,Sinjur,I.,Veber,I.,andZupan,B.,2008,Localitiesand sitesof Pulsatillavernalis intheJulianAlps:Hacquetia,v.7,p.47–69. Egli,B.R.,1991,Thespecialflora,ecologicalandedaphicconditions ofdolinesinthemountainofCrete:BotanicaChronika,v.10, p.325–335. Egli,B.R.,Gerstberger,P.,Greuter,W.,andRisse,H.,1990, Horstrissea dolinicola ,anewgenusandspeciesofumbels(Umbelliferae,Apiaceae) fromKriti(Greece):Willdenowia,v.19,no.2,p.389–399. Elliott,W.R.,2007,ZoogeographyandbiodiversityofMissouricavesand karst:JournalofCaveandKarstStudies,v.69,no.1,p.135–162. Favretto,D.,andPoldini,L.,1985,Thevegetationinthedolinasofthe karstregionnearTrieste(Italy):StudiaGeobotanica,v.5,p.5–18. Gargano,D.,Vecchio,G.,andBernardo,L.,2010,Plant-soilrelationships infragmentsofMediterraneansnow-beds:ecologicalandconservation implications:PlantEcology,v.207,no.1,p.175–189.doi:10.1007/ s11258-009-9663-7. Geiger,R.,1950,DasKlimaderbodennahenLuftschicht:EinLehrbuch derMikroklimatologie,thirdedition:Braunschweig,Friedr.Vieweg& Sohn,460p. Habel,J.C.,andAssmann,T.,eds.,2010,Relictspecies:Phylogeography andConservationBiology:Heidelberg,Springer,449p.doi:10.1007/ 978-3-540-92160-8. Horvat,I.,1953,Vegetacijaponikava:HrvatskiGeografskiGlasnik, v.14–15,p.1–25. Hoyk,E.,1999,Investigationofthevegetationandsoilinthedolinasof WesternMecsekMountains,SouthHungary:ActaCarsologica,v.28, no.1,p.105–113. Iverson,L.R.,andPrasad,A.M.,1998,Predictingabundanceof80 treespeciesfollowingclimatechangeintheEasternUnitedStates: EcologicalMonographs,v.68,no.4,p.465–485.doi:10.1890/00129615(1998)068[0465:PAOTSF]2.0.CO;2. Jakucs,L.,1977,Akarsztokmorfogenetika ja:Akarsztfejlo de svariencia i: Budapest,Akade miaiKiado ,284p. Judson,M.L.I.,2007,Anewandendangeredspeciesofthepseudoscorpiongenus Lagynochthonius fromacaveinVietnam,withnoteson chelalmorphologyandthecompositionoftheTyrannochthoniini (Arachnida,Chelonethi,Chthoniidae):Zootaxa,no.1627,p.53–68. Kiernan,K.,2002,ForestSinkholeManual:Tasmania,Hobart,Forest PracticesBoard,35p. Kira ly,G.,ed.,2009,U jMagyarFu ¨ve szko ¨nyv.Magyarorsza ghajta sos no ¨ve nyei.Hata rozo kulcsok:Jo svafo ,AggtelekiNemzetiParkIgazgato sa g,616p. Ko ¨hn,J.,andWaterstraat,A.,1990,Recentdistributionofglacialrelict MalacostracainthelakesofMecklenburg:AnnalesZoologiciFennici, v.27,p.237–240. Lausi,D.,1964,Vorla ¨ufigerU ¨ berblicku ¨berdieVegetationderTriester Karstdolinen:ActaBotanicaCroatica,v.4,p.65–71. Lazarevic ,P.,Lazarevic ,M.,Krivos ej,Z.,andStevanovic ,V.,2009,On thedistributionof Dracocephalumruyschiana (Lamiaceae)inthe BalkanPeninsula:PhytologiaBalcanica,v.15,no.2,p.175–179. Lewis,J.J.,andBowman,T.E.,2010,Thesubterraneanasellidsof Maryland:Descriptionof Caecidoteanordeni ,newspecies,andnew recordsof C.holsingeri and C.franzi (Crustacea:Malacostraca: Isopoda):JournalofCaveandKarstStudies,v.72,no.2,p.100–104. doi:10.4311/jcks2009lsc0092. Lova sz,Gy.,1971,AdatokazAbaligeti-karsztgeomorfolo giaie s hidrolo giaijellemze se hez:Fo ¨ldrajziE rtes to ,v.20,p.283–295. MacArthur,R.H.,andWilson,E.O.,1967,TheTheoryofIsland Biogeography:Princeton,PrincetonUniversityPress,203p. O ¨ zkan,K.,Gulsoy,S.,Mert,A.,Ozturk,M.,andMuys,B.,2010,Plant distribution-altitudeandlandformrelationshipsinkarsticsinkholesof MediterraneanregionofTurkey:JournalofEnvironmentalBiology, v.31,p.51–60. Pericin,C.,andHu ¨rlimann,H.,2001,Beobachtungenzurvertikalen VerteilungderMoosarteninderDolineSterna-FilariaimKarstgebiet vonBuje/BuieinIstrien(Kroatien):Bauhinia,v.15,p.91–96. Rockwood,L.L.,2006,IntroductiontoPopulationEcology:Oxford, BlackwellScientificPublications,339p. Sagarin,R.D.,Barry,J.P.,Gilman,S.E.,andBaxter,C.H.,1999,Climaterelatedchangeinanintertidalcommunityovershortandlongtime scales:EcologicalMonographs,v.69,no.4,p.465–490.doi:10.1890/ 0012-9615(1999)069[0465:CRCIAI]2.0.CO;2. Schindler,D.W.,Bayley,S.E.,Parker,B.R.,Beaty,K.G.,Cruikshank, D.R.,Fee,E.J.,Schindler,E.U.,andStainton,M.P.,1996,Theeffects ofclimaticwarmingonthepropertiesofboreallakesandstreamsat theExperimentalLakesArea,northwesternOntario:Limnologyand Oceanography,v.41,no.5,p.1004–1017. Sheldon,T.A.,Mandrak,N.E.,andLovejoy,N.R.,2008,Biogeographyof thedeepwatersculpin( Myoxocephalusthompsonii ),aNearcticglacial relict:CanadianJournalofZoology,v.86,p.108–115.doi:10.1139/ Z07-125. Simberloff,D.S.,andAbele,L.G.,1976,Islandbiogeographytheory andconservationpractice:Science,v.191,no.4224,p.285–286. doi:10.1126/science.191.4224.285. Simon,T.,2000,Amagyarorsza giede nyesflo rahata rozo ja:harasztok, vira gosno ¨ve nyek:Budapest,NemzetiTanko ¨nyvkiado ,845p. Szmorad,F.,1999,AdatokazAggteleki-karszte saGalyasa gflo ra ja hoz I.:Kitaibelia,v.4,no.1,p.77–82. Tichy ,L.,2002,JUICE,softwareforvegetationclassification:Journalof VegetationScience,v.13,no.3,p.451–453.doi:10.1111/j.1654-1103. 2002.tb02069.x. Trejo-Torres,J.C.,andAckerman,J.D.,2001,Biogeographyofthe Antillesbasedonaparsimonyanalysisoforchiddistributions: JournalofBiogeography,v.28,p.775–794.doi:10.1046/j.13652699.2001.00576.x. Vojtko ,A.,1997,U jadatokaTornai-karsztflo ra ja hoze svegeta cio ja hoz: Kitaibelia,v.2,no.2,p.248–249. Walther,G.-R.,Post,E.,Convey,P.,Menzel,A.,Parmesan,C.,Beebee, T.J.C.,Fromentin,J.-M.,Hoegh-Guldberg,O.,andBairlein,F.,2002, Ecologicalresponsestorecentclimatechange:Nature,v.416, p.389–437.doi:10.1038/416389a. Whiteman,C.D.,Haiden,T.,Pospichal,B.,Eisenbach,S.,andSteinacker, R.,2004,Minimumtemperatures,diurnaltemperatureranges,and temperatureinversioninlimestonesinkholesofdifferentsizes andshapes:JournalofAppliedMeteorology,v.43,p.1224–1236. doi:10.1175/1520-0450(2004)043 1224:MTDTRA 2.0.CO;2. Wolowski,K.,2003,Euglenophytesreportedfromkarstsink-holesinthe MalopolskaUpland(Poland,CentralEurope):AnnalesdeLimnologie-InternationalJournalofLimnology,v.39,no.4,p.333–346. doi:10.1051/limn/2003027. I MPORTANCEOFKARSTSINKHOLESINPRESERVINGRELICT,MOUNTAIN,ANDWET-WOODLANDPLANTSPECIESUNDERSUB -M EDITERRANEANCLIMATE: A CASESTUDYFROMSOUTHERN H UNGARY 134 N JournalofCaveandKarstStudies, April2012


Description
Articles:
http://dx.doi.org/10.4311/2011LSC0223 Cave Cricket Exit Counts:
Environmental Influences and Duration of Surveys / Floyd W.
Weckerly http://dx.doi.org/10.4311/2010MB0136R Involvement of
Bacteria in the Origin of a Newly Described Speleothem in the
Gypsum Cave of Grave Grubbo (Crotone, Italy) / Paola Cacchio,
Claudia Ercole, Rosita Contento, Giorgio Cappuccio, Maria
Preite Martinez, Maddalena Del Gallo, and Aldo Lepidi
http://dx.doi.org/10.4311/2011AN0219 The Prehistoric Cave Art
and Archaeology of Dunbar Cave, Montgomery County, Tennessee /
Jan F. Simek, Sarah A. Blankenship, Alan Cressler, Joseph C.
Douglas, Amy Wallace, Daniel Weinand, and Heather Welborn
http://dx.doi.org/10.4311/2010EX0167R Candidate Cave Entrances
on Mars / Glen E. Cushing http://dx.doi.org/10.4311/2011LSC0215
The First Subterranean Freshwater Planarians from North Africa,
with an Analysis of Adenodactyl Structure in the Genus
Dendrocoelum (Platyhelminthes, Tricladida, Dendrocoelidae) /
Abdul Halim Harrath, Ronald Sluys, Adnen Ghlala and Saleh
Alwasel http://dx.doi.org/10.4311/2010AN0148R1 Micro-Charcoal
Abundances in Stream Sediments from Buckeye Creek Cave, West
Virginia, USA / Gregory S. Springer, L. Nivanthi
Mihindukulasooriya, D. Matthew White, and Harold D. Rowe
http://dx.doi.org/10.4311/2010ES0178R Response of the Karst
Phreatic Zone to Flood Events in a Major River (Bohemian Karst,
Czech Republic) and its Implication for Cave Genesis / Helena
Vysok, Jir¡ Bruthans, Karel Z¡ k, and Jir¡ Mls
http://dx.doi.org/10.4311/2011jcks0193 A New Species of
Nicoletiidae (Insecta: Zygentoma) from Kartchner Caverns State
Park, Arizona / Luis Espinasa, Robert B. Pape, Alanna
Henneberry, and Christopher Kinnear
http://dx.doi.org/10.4311/2011es0222 Regionalization Based on
Water Chemistry and Physicochemical Traits in the Ring of
Cenotes, Yucatan, Mexico / Rosela Prez-Ceballos, Julia
Pacheco-vila, Jorge I. Eun-vila, and Hctor Hernndez-Arana
http://dx.doi.org/10.4311/2011jcks0197 Delineating Protection
Areas for Caves Using Contamination Vulnerability Mapping
Techniques: The Case of Herreras Cave, Asturias, Spain / A.I.
Marn, B. Andreo, M. Jimnez-Snchez, M.J. Domnguez-Cuesta,
and M. Melndez-Asensio http://dx.doi.org/10.4311/2011JCKS0213
Microbiological Activities in Moonmilk Monitored Using
Isothermal Microcalorimetry (Cave of Vers Chez Le Brandt,
Neuchatel, Switzerland) / Olivier Braissant, Saskia
Bindschedler, Alma U. Daniels, Eric P. Verrecchia, and
Guillaume Cailleau http://dx.doi.org/10.4311/2011LSC0216
Importance of Karst Sinkholes in Preserving Relict, Mountain,
and Wet-Woodland Plant Species under Sub-Mediterranean Climate:
A Case Study from Southern Hungary / Zoltn Btori, Lszl
Krmczi, Lszl Erds, Mrta Zalatnai, and Jnos Csiky.


printinsert_linkshareget_appmore_horiz

Download Options

close
Choose Size
Choose file type
Cite this item close

APA

Cras ut cursus ante, a fringilla nunc. Mauris lorem nunc, cursus sit amet enim ac, vehicula vestibulum mi. Mauris viverra nisl vel enim faucibus porta. Praesent sit amet ornare diam, non finibus nulla.

MLA

Cras efficitur magna et sapien varius, luctus ullamcorper dolor convallis. Orci varius natoque penatibus et magnis dis parturient montes, nascetur ridiculus mus. Fusce sit amet justo ut erat laoreet congue sed a ante.

CHICAGO

Phasellus ornare in augue eu imperdiet. Donec malesuada sapien ante, at vehicula orci tempor molestie. Proin vitae urna elit. Pellentesque vitae nisi et diam euismod malesuada aliquet non erat.

WIKIPEDIA

Nunc fringilla dolor ut dictum placerat. Proin ac neque rutrum, consectetur ligula id, laoreet ligula. Nulla lorem massa, consectetur vitae consequat in, lobortis at dolor. Nunc sed leo odio.