An Exploratory Non-Destructive Provenance Analysis of Two Middle Archaic Greenstone Pendants

17 Pages·2014·0.94 MB·English

Research Article
An Exploratory Non-Destructive Provenance Analysis of Two
Middle Archaic Greenstone Pendants from Little Salt Spring,
Florida, USA
MichaelF.Bonomo,1,*JustinP.Lowry,2RobertH.Tykot,3andJohnA.Gifford4
1DepartmentofGeological&EnvironmentalSciences,StanfordUniversity,Stanford,California
2DepartmentofSociology&Anthropology,GeorgeMasonUniversity,Fairfax,Virginia
3DepartmentofAnthropology,UniversityofSouthFlorida,Tampa,Florida
4DepartmentofAnthropology,UniversityofMiami,CoralGables,Florida
Correspondence Testexcavationsin2005and2006oftheLittleSaltSpringmortuarypondin
*Correspondingauthor; SouthwestFlorida(USA)yieldedtwoexoticstonependantsindirectlydatedto
E-mail:[email protected] the Middle Archaic (7000–5000 14C year B.P.). Hand-specimen petrographic
observation,combinedwithnon-destructiveenvironmentalscanningelectron
Received
microscopy, X-ray fluorescence spectroscopy, and whole-rock X-ray diffrac-
11March2011
Revised tion,identifiedthependantsas(1)amica-plagioclase-edeniteamphiboliteor
26November2013 schist and (2) an intermediate pyroxene- and/or amphibole-bearing grani-
Accepted toid.Provenancewasbroadlyconstrainedtoanumberofpotentialsourcesin
27November2013 the southernAppalachian Piedmontofthe United States (ata minimumdis-
tance of 650 km from the archaeological findspot) utilizing the United States
ScientificeditingbyDrewColeman
Geological Survey’s National Geologic Map Database lithologic search tool
PublishedonlineinWileyOnlineLibrary (GEOLEX)andMineralResourcesOn-LineSpatialDataGISdatabase.Dueto
(wileyonlinelibrary.com). theabsenceoflithologicmatcheswithinthestateofFlorida,weproposethat
thetwoartifactsmostlikelyarrivedatLittleSaltSpringthroughdown-the-line
doi10.1002/gea.21470
exchange of materials of prestige value with geologic origins in the southern
Appalachian Piedmont, arguing for the existence of small-scale long-distance
lithic exchange networks reaching into Archaic Florida. From a methodolog-
icalstandpoint,thisstudyillustratesthepotentialutilityofdatacollectedun-
der nonideal, non-destructive analytical conditions for deriving meaningful
archaeologicalinterpretations.(cid:2)C 2014WileyPeriodicals,Inc.
INTRODUCTION (e.g.,Nevin,Spoto,&Anglos,2012;Parish,Swihart,&Li,
2013),andsuchtechniquesarepreferred,ifnotrequired,
Theabilitytoidentifytherawmaterialsourcesofcultural
in instances where traditional destructive sample prepa-
heritage materials has developed into an important tool
ration is impermissible or where such preparation and
of archaeological inquiry. In addition to direct implica-
analysisistime-orcost-prohibitivegivenadesiredsam-
tions regarding resource procurement strategies (Parish,
plingvolumeorstrategy(Lundblad,Mills,&Hon,2008;
Swihart, & Li, 2013), provenance determinations often
Potts,2008;Artioli&Angelini,2011).
serve as proxies for a variety of human behaviors, in-
Combined approaches to archaeological sourcing are
teractions,andbeliefs,includingmobilityandterritorial-
commonly espoused, as the synthesis of multiple classes
ity(Burke,2006),directionalityandvolumeofexchange
of data carries a greater potential for making more se-
networks (Stoltman et al., 2005),and landscape percep-
cure source assignments (Neff, 2012; Nazaroff, Baysal,
tions(Vicensetal.,2010;Michelaki,Hancock,&Braun,
& C¸iftc¸i, 2013). With respect to lithic sourcing, a com-
2012).Theemploymentofnon-orminimally-destructive
binedapproachwillgenerallyincorporatebothgeochem-
technologies for archaeological analysis has flourished
icalandpetrographicdata.However,thenon-destructive
in the recent literature in response to technological ad-
potentialofcertaininstrumentaltechniquesforproviding
vancements and novel adaptations of existing methods
Geoarchaeology:AnInternationalJournal29(2014)121–137 Copyright(cid:2)C 2014WileyPeriodicals,Inc. 121
PROVENANCEOFLITTLESALTSPRINGGREENSTONES BONOMOETAL.
reliableandinterpretablegeochemicalorpetrographicin-
formation may not always be apparent due to a lack of
representationinthearchaeologicalliterature.Inparticu-
lar,whilescanningelectronmicroscopy(SEM)andX-ray
diffraction (XRD) are typically applied to archaeological
samplesthathaveundergonedestructivepreparationin-
volvingthepowdering,thinsectioning,orsputtercoating
ofsamples(e.g.,Sˇegvic´etal.,2012),thesetwobasictech-
niquesarealsocapableofbeingappliednon-destructively
intheanalysisofculturalheritagematerials.
Here we present the results of a combined non-
destructivegeochemicalandmineralogicalstudyaimedat
determiningtheprovenanceoftwoarchaeologicalgreen-
stone pendants from the Paleoindian- to Archaic-period
LittleSaltSpringsiteinFlorida,USA.Ouruseoftheterm
“greenstone”hereincorrespondstoaloosearchaeological
definitionratherthanastrictgeologicaldefinitioncarry-
ing mineralogical or petrogenetic significance. The pen-
dants reported here represent two of only three archae- Figure1 StudyareainthesoutheasternUnitedStates,showingFlorida
ological greenstones discovered among Florida’s Archaic archaeologicalsitesmentionedinthetext.LSS=LittleSaltSpring(8SO18);
sites to date, and are among Florida’s oldest ceremonial BP=BrickellPoint(8DA12);HI=HogIsland/GraveYardIsland(8CI220);
GM=GoodmanMound.FL=Florida;AL=Alabama;GA=Georgia;SC=
artifacts to be analyzed for provenance (the third pen-
SouthCarolina;NC=NorthCarolina.
dant, from the Republic Groves site in Hardee County,
Florida [8HR4], is, at the time of writing, on display in
anexhibitattheUniversityofFlorida’sMuseumofNatu-
FloridaAquariumrecoveredastonependant(designated
ralHistoryinGainesville).Thiscasestudyservesnotonly W24S04A04)froma4×4msurveysquarein7.6mof
as an interpretation of early practices of exotic stone re-
water.Asecond,smallerstonependant(W24N04WA03)
source acquisition among Florida’s Archaic peoples, but
was recovered in June of 2006 on the east basin slope
on a broader scale as a demonstration of the potential
approximately 10 m north of the findspot of the first
utility of non-destructive SEM and XRD methodologies
pendantinsurveysquareW24N04W.Artifactsrecovered
inthefieldofarchaeology.
fromthisportionoftheupperbasin,aswellassomedis-
articulatedandcommingledhumanskeletalremains,are
believed to represent Middle Archaic burials that have
ARCHAEOLOGICAL CONTEXT naturallyerodedfromtheshallowuppermostslopeofthe
basin and, over the millennia, migrated downslope and
Little Salt Spring (8SO18) in southern Sarasota County,
overthedrop-offundertheinfluenceofgravity.Thetwo
Florida, USA (Figure 1), is a flooded spring-fed sink-
pendants are indirectly dated to the Middle Archaic pe-
hole once used as a mortuary pond by Archaic-period
riod by means of 14C dating of other artifacts from the
groups(Clausenetal.,1979;Wentz&Gifford,2007);ad-
same survey square in which the smaller pendant was
ditionally, it may have functioned as a freshwater oasis
found.
andnaturalanimaltrapduringthePaleoindianandearly
Archaic periods (Gifford, 1993). Well-preserved organic
artifactsmadeofanimalboneandantler,aswellaseco- MATERIALS
facts(wood,charcoal,andpeat)haveyielded14Cagesfor
W24S04A04
theoccupationsatLittleSaltSpringrangingfrom12,000–
9000 14C year B.P. for Paleoindian cultural remains and Thefirstofthependants,W24S04A04(subsequentlyre-
6800–5200 14C year B.P. for Archaic artifacts (Clausen ferredtoasA04;Figure2A)islikelyanalteredamphibo-
et al., 1979). Approximately 100 more radiocarbon de- liteoramphiboleschist.ThemassofA04wasmeasuredat
terminationssince1992indicatearoughlysimilarrange 12.0 g with an average specific gravity of G = 2.84 over
ofradiocarbondatesfrom12,500–600014CyearB.P. eight replicate measurements. It is polished, biconically
Duringunderwatertestexcavationsontheeastslopeof drilled, and carved into a parabolic curve extending dis-
the upper basin of Little Salt Spring in 2005, University tallyfromthedrilledend.Itsmaximumlengthis49mm
of Miami research divers and volunteer divers from the and28mminwidthacrossitsproximal(drilled)end.
122 Geoarchaeology:AnInternationalJournal29(2014)121–137 Copyright(cid:2)C 2014WileyPeriodicals,Inc.
BONOMOETAL. PROVENANCEOFLITTLESALTSPRINGGREENSTONES
Figure2 (A)PendantW24S04A04(verticalaxisapproximately49.0mminlength).(B)PendantW24N04WA03(verticalaxisapproximately26.5–27.0mm
inlength).
W24N04WA03 archaeological literature involve destructive sam-
plingofthematerialsunderstudy(eitherfragmentation
The second pendant, tagged as W24N04WA03 and sub-
orthinsectioning)necessitatedbysmallsamplechamber
sequently referred to as A03 (Figure 2B), was identified
dimensions relative to a larger sample (Janssens et al.,
asacoarse-grainedpyroxene-and/oramphibole-bearing
2000; Mantler & Schreiner, 2000). The non-destructive
granitoid.Itisroughlyrectangularinshape,withalength
capabilitiesofSEMinstrumentsarethereforedependent
rangingfrom26.5to27mm.ThemassofA03wasmea-
on instrument-specific sample chamber size limitations,
suredat1.8gwithanaveragespecificgravityofG=2.57
andalsoontheavailabilityofadetectorcapableofoper-
over five replicate measurements. The pendant is also
ating in a gaseous atmosphere as opposed to a vacuum.
polished and biconically drilled at one end. Both thick-
SuchvariantsoftheSEMaretermedESEMs,andunlike
nessandwidthofthestoneincreaseinthedirectiondis-
traditional SEM instruments, are capable of analyzing
taltothedrillhole,withawidthof8.5mmandthickness
materials without the need for contaminative sputter
of 2 mm near the drill hole and a width of 10 mm and
coating or other forms of insulation or pretreatment
thicknessof3mmatthedistalend.
(Danilatos, 1991). ESEM-EDS analyses were performed
on both archaeological pendants at the University of
METHODOLOGY Miami’s Center for Advanced Microscopy (UMCAM) on
the Coral Gables, Florida, campus on an FEI XL-30 field
GeneralMethods emission environmental scanning electron microscope.
ESEM operation was conducted in an atmosphere at
Due to the rarity and value of these artifacts, only non-
1.4 Torr, and in situ qualitative chemical analyses were
destructive analytical techniques were employed. Pen-
obtained via EDS operating with an accelerating voltage
dants A03 and A04 were analyzed using a combination
of20.0kVandabeamdiameterof3μm.
of environmental SEM coupled with energy-dispersive
X-ray fluorescence spectroscopy (ESEM-EDS),
laboratory-based and portable XRF (XRF and pXRF),
X-rayFluorescenceandPortableXRF
whole-rock XRD, macroscopic petrographic observation,
andwhole-rockspecificgravityanalysis. XRF has been widely utilized in the archaeological field
to provide both qualitative and quantitative measures
oftheelementalcompositionsofmaterialsandpotential
EnvironmentalScanningElectronMicroscopy
sources (Shotton & Hendry, 1979; for a recent example,
TheapplicationofSEMtothesourcingofgeologicalraw seeNazaroff,Baysal,&C¸iftc¸i,2013).XRFtheoryiswell-
materials has long been realized (e.g., Freestone & Mid- established in the literature, and readers are referred to
dleton, 1987). However, with respect to lithics, the thedescriptionsofthetheoryandprinciplesofXRFpre-
majority of SEM methodologies detailed in the sentedbyBertin(1970)andPotts(1987).
Geoarchaeology:AnInternationalJournal29(2014)121–137 Copyright(cid:2)C 2014WileyPeriodicals,Inc. 123
PROVENANCEOFLITTLESALTSPRINGGREENSTONES BONOMOETAL.
Multiple independent qualitative XRF analyses were entdspacingsbetweenparticularcrystallographicplanes,
performed on A04: one at the State University of New allowing minerals to be identified from a diffractogram
York at Albany’s (SUNY—Albany) Ion Beam Laboratory (Moore & Reynolds, 1997). Mineralogical identification
and a second at the University of South Florida’s (USF) via XRD is often viewed as “unreliable” for objects
LaboratoryforArchaeologicalSciences.Analysesofboth with curved, complex, or otherwise nonflat surfaces
artifacts were undertaken with no destructive sample (Zhigachev, 2013), usually necessitating the destructive
preparation.TheXRFinstrumentatSUNY—Albany’sIon preparation of pressed powders, pellets, or slides (read-
Beam Laboratory has an estimated resolution of 165 eV ers are referred to Moore & Reynolds, 1997, for more
for Fe at 6.40 kV. The sample chamber was not evac- detailed information on the general theory of XRD and
uated. Characteristic X-rays from the sample were de- common sample preparation techniques). As the accu-
tectedbyaSi(Li)detector,counted,andsortedbyamul- racy of the observed d spacing is highly dependent on
tichannel analyzer. The analytical procedure was based maintainingaprecisegeometricorientationbetweenthe
on previous experimental XRF work (Kuhn & Lanford, incident source X-rays, the sample plane, and the de-
1987;Stevenson,Klimkiewicz,&Scheetz,1990;Hermes tector, it is inevitable that peaks produced in any non-
&Ritchie,1997;Williams-Thorpe,Potts,&Webb,1999). destructive whole-rock scans of curved surfaces will not
Two XRF analyses were performed on A04 at the Ion be perfect matches with published powder diffraction
BeamLaboratorywithascantimeof240minuteseach, data,evenwiththeapplicationofcorrectionprocedures.
onewithaSntargetandtheotherwithaTatarget.The However,whileuncommon,precedentdoesexistforthe
Sntargetyieldedbetterresults,asseveralofthetraceel- careful archaeological interpretation of non-destructive
ements identified in the scan would have been indistin- whole-rock XRD diffractograms, with Steponaitis et al.
guishableifanalyzedwiththeTatargetalone.Elemental (2011: 86) asserting that any “interpretive difficulties...
peakswereidentifiedwiththecomputerprogramsAnal- [are]easilymitigatedwithadetailedvisualexamination
ysisandAXIL.XRFanalysisofA03wasnotconductedat ofthediffractionpatterns,coupledwiththeinformation
SUNY—Albany. gainedfromthehand-samplepetrology.”
TheanalysisofbothA04andA03attheLaboratoryfor Non-destructive XRDanalysiswasperformedonpen-
ArchaeologicalSciencesatUSFutilizedaBrukerTracerIII dantA04onlyattheUniversityofGeorgia’sDepartment
serieshandheldXRF(pXRF)analyzerwithsettings(40.00 ofGeologyonaBrukerD8-Advancediffractometer,with
kV accelerating voltage, 10.00 μA current, no vacuum) theresultsprocessedusingDIFFRAC.SUITETOPASsoft-
and filters aimed at quantitatively measuring the pres- ware; the small size of A03 relative to the XRD sample
enceofFeandtraceelementsRb,Sr,Y,Zr,andNb.Spec- holderdidnotallowtheartifacttobeadequatelyoriented
trawerecollectedfor300secondseachandrawphoton for analysis. For the XRD analysis of A04, no sample
countswereconvertedtopart-per-millionconcentrations preparation was undertaken other than the surficial ap-
utilizing the five UMC calibration program developed at plicationofapowderedZnOanalyticalstandard(National
theUniversityofMissouri—Columbiaanddistributedby Institute of Standards and Technology Standard Refer-
BrukerCorporation.Therelativeproportionsofelements enceMaterial674a).Thestandardwassmoothedontoa
suchasK,Ca,andTiwerealsorecorded,thoughcalibra- portionofthesurfaceoftheartifactinanattempttocor-
tionswerenotavailablefortheseelementsonthisinstru- rectforsampledisplacementcausedbythecurvedsurface
mentusingthisparticularconfiguration.Forgeneralcon- oftheartifactextendinginsidethediffractometer’sfocus-
siderationsrelatingtotheapplications(andlimitations)of ing circle and was removed by rinsing in water follow-
pXRF,readersarereferredtoPottsandWest(2008),and ingtheanalysis.Theuseofinternalstandards,wherethe
ShugarandMass(2012). standard is mixed with the powdered sample, is a com-
monanalyticaltechniqueinXRD(e.g.,Hurst,Schroeder,
& Styron, 1997; S´rodon´ et al., 2001). The powdering of
X-rayDiffraction
A04, however, was not an option; hence, the standard
XRD is a spectroscopic technique that measures the dis- was applied externally and smoothed onto the surface
tance,d,betweenlatticeplanesinacrystallinesampleby of the artifact as best as possible. Similar methods are
relating this spacing between planes to the wavelength presentedinSteponaitisetal.(2011).Applicationofthe
(λ)andangle(θ)oftheincidentX-raysaccordingtothe standard for XRD was conducted only after the ESEM-
Braggequation EDS and XRF chemical analyses were performed in or-
dertoavoidspectralartifactsresultingfromanypotential
nλ=2dsinθ (1)
surficialcontaminationbyZn.
wherenisanintegerrepresentingthenumberofwave- Two XRD analyses were performed on A04 using
lengths. Different minerals have characteristically differ- CoKα1 radiation with a wavelength of 1.79 A˚, an
124 Geoarchaeology:AnInternationalJournal29(2014)121–137 Copyright(cid:2)C 2014WileyPeriodicals,Inc.
BONOMOETAL. PROVENANCEOFLITTLESALTSPRINGGREENSTONES
1.62A˚.Toobtaincorrectionsoutsideoftherangecovered
bythestandards,alinearregressionwasperformedusing
all data points from the standard (the solid gray line in
Figure 3B), and a segment was formed between (1) the
calculatedcorrectionat2θ =0o (atadcorrectionofap-
proximately0.0781A˚,ascalculatedfromtheregression;
thepointlabeled“1”inFigure3B)andthefirstZnOpeak
in the scan (at 37.57o and 0.0425 A˚, point “2”), and (2)
between the last ZnO peak in the scan (at 67.27o and
0.0121A˚,point“6”)andtheregression-calculateddcor-
rectionattheendofthescan(75.00oand0.0021A˚,point
“7”).Assuch,moreconfidencecanbeplacedintheaccu-
racy of the correctionsapplied withinthe rangecovered
by the ZnO standard; the least confidence is placed on
thosevaluesapproachingd=14A˚.
Forty minerals with known green color variants
(e.g., several pyroxene, amphibole, serpentine, and
Figure3 (A)XRDdiffractogramofA04.Axesareinunitsof2θangle(in chlorite-group minerals), or otherwise common
degrees,x-axis)andcounts(y-axis).Thedarker,higherintensitypattern
rock-forming minerals (e.g., quartz, feldspars, and
wasproducedpriortotheapplicationoftheNISTSRM674aZnOstandard.
micas), were selected from the Mineralogical Soci-
Thelighter,lowerintensitypatternwasproducedaftersurficialapplication
ofthestandard.(B)dSpacingcorrectionsappliedtopeaksproducedin ety of America’s (MSA) Crystal Structure Database
theXRDdiffractogramofA04.Axesareinunitsof2θangle(indegrees, (http://www.minsocam.org/MSA/crystal database.html;
x-axis,alignedwiththex-axisinFigure3A)andcorrection,d*.Numbered Downs & Hall-Wallace, 2003) for comparison with the
datapointscorrespondtoNISTSRM674aZnOstandardpeaksidentified corrected peak d spacings obtained from the diffrac-
inthediffractogram:2—(100)plane;3—(002);4—(101);5—(102); togramofpendantA04(seeTableIIIforthecompletelist
6—(110).Points1and7correspondtothebeginningandendofthe
of minerals). Automated search-match algorithms were
scan,andarenotZnOpeaks.Linearequationsoftheformd*=mx2θ+b
not utilized as such algorithms rely on peak intensity,
representthecorrectionequationsappliedtopeaksidentifiedalongeach
which was considered unreliable in the artifact scan, in
segmentofthescan(i.e.,betweenZnOpeaks),andwereusedtoobtain
thedspacingsgiveninTableII.Verticaltielinesconnectingpoints2–6in addition to d spacings. Mineral phases were assigned to
Figure3BtopeaksinthediffractograminFigure3AidentifytheZnOpeaks peaks on the arbitrary basis of a “good” match consti-
inthescan;thetielineatpoint7indicatestheendofthescan. tuting a corrected d spacing occurring within 0.0009 A˚
of a published d spacing for that mineral, with the
potentialformultiplephasestobepreliminarilyassigned
accelerating voltage of 40 kV and a current of 35 mA, toasinglepeak.Inaddition,the“best”matchormatches
atascanspeedof1o2θ perminutewithastepincrement were defined as the phases with the closest published d
of0.02ofrom5.00oto75.00o2θ.Resultsobtainedonthe spacing relative to a given corrected d spacing from the
instrumentareprecisetoatleastfourdecimalplaces.One diffractogram. The 40 individual mineral phases were
scanwasrunpriortotheapplicationoftheZnOstandard then grouped into general mineral groups (e.g., micas,
andtheotherafterthestandardwasapplied(Figure3A). pyroxenes, amphiboles, etc.) for subsequent determina-
In this manner, the ZnO peaks were easily identified in tion of which mineral groups could best account for the
thescanandtheappropriatecorrectionswereappliedto remainingnon-ZnOpeaksinthediffractogram.
thepeakdspacingsfromtheartifacttocorrectforsample
displacement created by the irregular shape of the pen-
SpecificGravity
dantrelativetothesampleholder(TableI).Separatecor-
rections were calculated for each segment of the scan, Specific gravity (G) is a unitless ratio of a material’s
that is, between each set of adjacent data points for the mass to that of an equal volume of water (Klein, 2002:
ZnOstandardbymeansofalinearregression(Figure3B). 33).SpecificgravitymeasurementsofA03andA04were
It should be noted that peaks for the ZnO standard conducted at the University of Georgia through the hy-
covered only the range of d from approximately 1.62– drostatic weighing method with a digital apparatus (see
2.61 A˚, and that 25 peaks were identified within the Rapp, 2009: 26). Values were calculated by measuring
range∼14>d>2.61A˚ (thoughonlyfourpeakswerede- eachartifact’smassinair(m )followedbymeasuringits
a
tectedatd>5A˚),whiletwopeakswereidentifiedatd< mass as suspended in water (m ) at room temperature
w
Geoarchaeology:AnInternationalJournal29(2014)121–137 Copyright(cid:2)C 2014WileyPeriodicals,Inc. 125
PROVENANCEOFLITTLESALTSPRINGGREENSTONES BONOMOETAL.
TableI Certified,observed,andcorrectedXRDdspacingsforNISTSRM674astandardappliedtoartifactA04.
NISTSRM674a(plane) 2θCuKα(o)certaλ=1.54A˚ 2θCoKα(o)certbλ=1.79A˚ 2θCoKα(o)obscλ=1.79A˚ dobs(A˚)d dcorrected(A˚)e
100 31.70 37.02 37.57 2.7778 2.8204
002 34.36 40.16 40.71 2.5717 2.6079
101 36.18 42.32 42.88 2.4469 2.4807
102 47.48 55.80 56.32 1.8953 1.9134
110 56.52 66.78 67.27 1.6148 1.6269
103 62.80 74.54 Notobserved Notobserved Notobserved
200 66.30 78.92 Notobserved Notobserved Notobserved
112 67.88 79.42 Notobserved Notobserved Notobserved
aNISTSRM674acertifiedpeakpositionsforthe(100),(002),(101),(102),(110),(103),(200),and(112)planesforCuKαradiationwithawavelength
λ=1.54A˚.
bCertifiedpeakpositions,ascalculatedforCoKαradiationwithawavelengthλ=1.79A˚.
cObservedNISTSRM674apeakpositions.
dObservedNISTSRM674adspacings.
eCorrectedNISTSRM674adspacings(correctedaccordingtotheequationsinFigure3B).
andapplyingtheformula
m
G = a (2)
ma−mw
RESULTS AND INTERPRETATIONS
W24N04WA03
Elemental ESEM-EDS analysis of A03 indicates a bulk-
rock chemistry consisting largely of Si and Al with
lesser K and minor amounts of Ca, Na, Fe, and Ti
(Figure 4A). Initially thought to be a pyroxene-
plagioclase diabase/gabbro or diorite in hand sam-
ple, the presence of measurable potassium revealed
through EDS is likely indicative of an interme-
diate granitoid as opposed to diabase/gabbro/diorite Figure4 (A)SEM-EDSspectrumofA03,operatingwithanaccelerating
(Nockolds, 1954). Further interpretation of the EDS voltageof20.0kV.Thex-axiscoversthespectrumofenergiesfrom0to
spectrum indicates that the feldspar component of 8kV,andthey-axisisinrelativeintensityKαX-raysforO,Na,Al,Si,K,Ca,
A03 is a combination of K-feldspar [KAlSi O ] and Ti,andFearelabeled.(B)SEM-EDSspectrumofA04,operatingwithan
3 8
acceleratingvoltageof20.0kV.Thex-axiscoversthespectrumofenergies
an intermediate plagioclase feldspar (intermediate be-
tween the NaAlSi O [albite] and CaAl Si O [anor- from0to8kV,andthey-axisisinrelativeintensityKαX-raysforO,Na,Al,
3 8 2 2 8 Si,K,andCaarelabeled.
thite] end members) and that contributions of cal-
cium and sodium are made from plagioclase and the
pyroxene mineral augite [(Ca,Na)(Mg,Fe,Al)(Si,Al) O ] [Ca (PO ) (F,Cl,OH)]inclusionsnotvisibleinhandsam-
2 6 5 4 3
or possibly from combinations of plagioclase, pyrox- plewerealsodetectedwithEDSinelementmapswhere
ene, and amphibole minerals such as hornblende highconcentrationsofCaandPoverlap(Figure5D–F).
[(Ca,Na) (Mg,Fe,Al) Si (Si,Al) O (OH) ]. The pres- Data obtained from A03 are consistent with a litho-
2–3 5 6 2 22 2
ence of Mg-bearing pyroxene and/or amphibole min- logic interpretation of an intermediate, potentially al-
erals likely accounts for the greenish coloration of the tered,pyroxene-and/oramphibole-bearinggranitoid.By
rockobservableinhandsample.Accessorymineralscon- definition, “granitoids” are a family of rocks of gen-
sist primarily of minor amounts (less than 3% by vol- eral granitic composition, which include “true” gran-
ume) of the Fe-Ti oxide mineral ilmenite [FeTiO ] (Fig- ites, alkali feldspar granites, granodiorites, and tonalities
3
ure5A–C),thoughthebulk-rockFesignaturesmayalso (Streckeisen, 1973, 1976). Chemically, granitoid rocks
be partially derived from amphiboles. Tabular apatite display a general trend in which SiO > Al O > K O
2 2 3 2
126 Geoarchaeology:AnInternationalJournal29(2014)121–137 Copyright(cid:2)C 2014WileyPeriodicals,Inc.
BONOMOETAL. PROVENANCEOFLITTLESALTSPRINGGREENSTONES
Figure 5 (A)StereoscopicmicroscopeimageofanilmenitecrystalinA03.(B)SEM-EDSspectrumofanilmenitecrystalinA03,operatingwithan
acceleratingvoltageof30.0kV.(C)SEM-EDSelementmapofanilmenitecrystalinA03,showingKαX-raysforTi(left)andFe(right).(D)Secondaryelectron
imageofanapatitecrystalinA03.(E)SEM-EDSspectrumofanapatitecrystalinA03,operatingwithanacceleratingvoltageof20.0kV.(F)SEM-EDS
elementmapofanapatitecrystalinA03,showingKαX-raysforCa(left)andP(right).
> Na O ≥ CaO (Clarke, 1992). Although not reliably concentrationofthetraceelementsRb,Sr,Y,Zr,andNb,
2
quantifiable in terms of oxide weight percentages, the aswellasshowingasubstantialFepeakandanapprecia-
elementalEDSresultsfromA03seemtoreflectthistrend. bleTipeak(Figure6A).
While classified primarily according to the relative pro- SpecificgravitymeasurementsofA03(G=2.57)agree
portionsofquartzandfeldspars,granitoids,suchasthose with a granitoid identification. It is important to note,
in the southern Appalachian Mountains, also contain however, that the specific gravity measurement of G =
minor or accessory amphiboles, pyroxenes, and heavy 2.57forA03representsthemostcommonresultobtained
minerals such as rutile and ilmenite (McSween, Speer, over multiple replicate measurements, and that minor
& Fullagar, 1991; Tollo et al., 2004). The lack of ob- variationsof0.1gineitherorbothofthemassinairand
servable biotite mica in A03, a common mafic compo- massinwaterresultedinalargerangeofcalculatedspe-
nent of granitoid rocks, may be a reflection of (1) rela- cificgravitiesfromG=2.38–2.71(theeffectofthesemi-
tive enrichment of Fe over Mg in the rock (Frost et al., norvariationsareenhancedbythelowmassofA03).Us-
2001) or (2) the hydrothermal alteration and replace- ingthespecificgravitiesofquartz(G=2.65),K-feldspars
ment of biotite with fine-grained chlorite group miner- (G = 2.54–2.62), plagioclase feldspars (G = 2.62–2.76),
als [(Mg,Fe) (Si,Al) O (OH) ·(Mg,Fe) (OH) ] (Fiebig & augite (G = 3.2–3.4), hornblende (G = 3.0–3.4), apatite
3 4 10 2 3 6
Hoefs,2002),theMgcomponentofwhichwouldbeun- (G = 3.16–3.22), and ilmenite (G = 4.7), and assuming
detectable by EDS in a nonvacuum because of the di- (1) the upper estimate of G = 2.71 for A03 and (2) the
minished ability of the low-energy Mg X-rays to travel lowerestimatesofspecificgravityforaugite,hornblende,
throughanatmosphereenroutetothedetector(Stepon- apatite, and the feldspars are applicable, then calculated
aitisetal.,2011).ThepXRFanalysisofA03measuresthe whole-rock specific gravities of G = 2.71 or lower are
Geoarchaeology:AnInternationalJournal29(2014)121–137 Copyright(cid:2)C 2014WileyPeriodicals,Inc. 127
PROVENANCEOFLITTLESALTSPRINGGREENSTONES BONOMOETAL.
Figure6 (A)pXRFspectrumofA03(darker)andA04(lighter),operatingat40.0kVand10.00μAfor300seconds.Thex-axiscoverstheenergyspectrum
from1.20to18.00kV,andthey-axisisinrelativecounts.Kα1spectrallinesforK,Ti,Cr,Fe,Cu,Zn,Ga,Rb,Sr,Y,Zr,andNbarelabeled,aswellastheKβ1
spectrallinesforFeandZr(denotedbyan*).Theinsetatupperleftpresentsthecalibratedconcentrations(ppm)ofselectedelements.(B)SUNY—Albany
IonBeamLaboratoryXRFspectrumofA04,showingαwavepeaksforboththeSn(darker)andTa(lighter)targets.Peakscorrespondingtotheαwaves
ofRb,Sr,Zr,andBaarelabeled.
TableII Calculatedwhole-rockspecificgravitiesfortheoreticalgranitoids green coloration to greenstones and amphibolites, for
containingthemineralassemblageidentifiedinartifactA03. example, the chlorite group minerals, serpentine group
minerals [Mg Si O (OH) ], or amphibole minerals such
Pyroxene 3 2 5 4
Modalcomposition(%) Tonalite Granodiorite Granite granite as hornblende, actinolite [Ca2(Mg,Fe)5Si8O22(OH)2], or
edenite [NaCa Mg AlSi O (OH) ], are largely magne-
Quartz 25 30 35 20 2 5 7 22 2
sian in composition; as previously stated, this is prob-
Plagioclase 55 35 20 35
K-feldspar 5 25 35 40 lematic for EDS detection in a nonvacuum. Thus, the
Hornblende 10 4 6 1 requirement to operate the ESEM in an atmosphere
Augite 3 4 2 2 did not allow for the detection of Mg, even if it may
Ilmenite 1.8 1.8 1.8 1.8 have been present in the sample. Visual similarities
Apatite 0.2 0.2 0.2 0.2
between A04 and a predominantly plagioclase- and
Calc.specificgravity 2.71 2.69 2.68 2.65
edenite-bearing amphibolite/amphibole schist from the
ModalmineralogiesareroughlybasedonthosepresentedinMa¨kitieetal. BuckCreekmafic/ultramaficcomplexintheeasternBlue
(1999),excludingmicaphases. Ridge province of western North Carolina suggests that
edeniteandplagioclasemayconstitutetheprimarymin-
obtainable using modal mineralogies reasonably similar eralogyofA04(Figure7).TheESEM-EDSelementalpro-
tothosepresentedinMa¨kitieetal.(1999)forseveralpos- filemaybetakentosupportthisdetermination.Edenite,
siblegranitoidcompositions(TableII). unlikeothercommonamphibolesassociatedwithgreen-
stones and amphibolites, lacks stoichiometric Fe. Eden-
ite also contains both stoichiometric Na and Ca, which,
W24S04A04
unlike Fe, are identifiable in the spectrum. Plagioclase
Elemental ESEM-EDS analysis of A04 reveals a bulk- feldsparwouldalsocontributeNaand/orCatothechem-
rock chemistry predominantly consisting of Al and Si icalprofile.
with lesser amounts of K, Na, and Ca (Figure 4B). Pre- This interpretation alone, however, does not account
liminaryhand-samplevisualexaminationidentifiedA04 for the high-intensity Al peak and moderate K peak in
as potentially either a true greenstone or an amphibo- the spectrum. In addition to the dominant green min-
lite. The minerals typically responsible for imparting the eral phase or phases, patches of what appear to be a
128 Geoarchaeology:AnInternationalJournal29(2014)121–137 Copyright(cid:2)C 2014WileyPeriodicals,Inc.
BONOMOETAL. PROVENANCEOFLITTLESALTSPRINGGREENSTONES
TableIII IdentifiedpeaksintheX-raydiffractogramofA04
dcorrected BestMatch Matcheswithin0.0009A˚
14.2243 chla,b
9.5888 micac
9.3097 zeob,d
5.2903 zeob
4.9308
4.7563 pyxe ampf,chl
4.4018 pyx mica,KASg,chl
4.3559 mica
4.2440 mica chl
4.0182 pyx amp
3.8546 KAS,plh,chl,nei
3.7648 clayj mica,zeo,KAS,pl,kfsk
3.7105 zeo
3.6491 pyx amp
3.4773 zeo,kfs
3.4059 zeo pl Figure7 Visualcomparisonofaplagioclase-edeniteamphibolite/schist
3.3757 mica KAS,pl,clay fromtheBuckCreekmafic/ultramaficcomplexinClayCounty,NorthCar-
3.3590 amp mica,qzl olina,USA(left)andA04(right).
3.3236 zeo mica,pyx,amp,chl,kfs
3.2605 pl,clay kfs
3.1991 mica,pyx,pl whitemicaphase(e.g.,muscovite[KAl (AlSi O )(OH) ]
2 3 10 2
3.1539 pyx,KAS,chl,clay
ormargarite[CaAl (Al Si ) (OH) ])arelikelycontribut-
2.9883 pyx,kfs mica,zeo,chl 2 2 2 10 2
2.9196 pyx mica,pl ingsubstantialAland/orKtotheelementalprofileofA04
2.8638 chl mica,pyx,pl (Figure8C).AdditionalaccessorymineralsinA04include
2.8203 ZnOm
2.7995 kfs mica,pyx,zeo,srpn reddish-coloredTi-bearingmineralinclusionsinterpreted
2.6920 KAS zeo asrutile[TiO ]onthebasisofEDSelementmapslacking
2
2.6078 ZnO
significantconcentrationsofFeinassociationwiththeTi
2.5810 mica,kfs pyx,zeo,KAS,amp,clay,ne
2.5612 amp mica,pyx,zeo,pl,chl,clay,srp (Figure 8A and B). In some areas, the rutile inclusions
2.5289 pl mica,pyx,zeo,amp,chl,clay,kfs,srp gradefroma deep redtoa reddish-brownand finallyto
2.4807 ZnO
2.4661 amp mica,pyx,zeo,KAS,chl,qz a yellowish-brown phase, possibly indicating differential
2.4404 mica,zeo,KAS,pl,amp,chl,ne,srp progression in the stages of alteration or breakdown of
2.4293 KAS,pl mica,pyx,zeo,amp,chl,clay
the rutile to leucoxene, a common yellow/brown alter-
2.3777 ne pyx,zeo,KAS,pl,chl,kfs
2.3571 KAS mica,pyx,zeo,pl,clay ation product of Ti-bearing minerals (Allen, 1956; Bai-
2.3443 ne mica,pyx,zeo,KAS,amp,clay ley et al., 1956; Karkhanavala & Momin, 1959; Guilbert
2.2499 mica,pyx zeo,KAS,amp,chl,clay,qz
2.1991 clay mica,pyx,zeo,KAS,pl,amp,chl,srp & Park, 2007). Also of note are several small embed-
2.1498 srp mica,pyx,zeo,KAS,amp,clay,ne,srp ded monazite [(LREE)PO ] crystals revealed by EDS to
4
2.0025 mica pyx,KAS,pl,amp,chl,clay,kfs
contain significant amounts of the light rare earth ele-
1.9475 pyx mica,zeo,KAS,pl,amp,chl,clay,srp
1.9241 mica,pyx zeo,KAS,pl,amp,kfs,ne,srp ments (LREE) Ce and Nd in association with high con-
1.9134 ZnO
centrationsofPandminoramountsofCa(Figure8Dand
1.6817 zeo,KAS,pl,clay mica,pyx,amp,chl,kfs,srp,qz
1.6269 ZnO E).Whilemonaziteoftencontainsminorconcentrations
1.6154 pyx mica,zeo,KAS,pl,amp,chl,clay,ne of Ca (Flinter, Butler, & Harral, 1963; Zhu & O’Nions,
1.6089 mica,pyx zeo,KAS,pl,amp,clay,ne
1999),itisalsopossiblethatthesecrystalsareLREE-rich
achloritegroupminerals(chlorite,clinochlore)
bbestmatchforagivenpeak,butnotwithin0.0009A˚ apatites.
cmicagroupminerals(muscovite,paragonite,margarite,phengite) The geochemical signature and mineralogical assem-
dzeolitegroupminerals(heulandite,stilbite) blage of A04 is highly suggestive of a lithology simi-
epyroxenegroupminerals(augite,enstatite,diopside,jadeite,omphacite, lar to the rocks of the Buck Creek complex, in which
hypersthene,hedenbergite) Berger et al. (2001) identify edenite-margarite schists
famphibolegroupminerals(tremolite,actinolite,edenite,anthophyllite)
(hydrothermally altered metatroctolites) with abundant
galuminosilicategroupminerals(kyanite,andalusite,sillimanite)
modalplagioclaseandstronglyelevatedLREEsignatures.
hplagioclasegroupminerals(albite,anorthite)
inepheline Edenite-bearingassemblagesatBuckCreekandotherre-
jclaygroupminerals(talc,kaolinite,pyrophyllite) lated mafic-ultramafic complexes in western North Car-
kK-feldspargroupminerals(orthoclase,sanidine,microcline) olina (e.g., Tathams Creek and Lake Chatuge) typically
lquartz containemeraldorgrassy-greenedeniteamphibole(50–
mNISTSRM674a 70 modal% in combination with magnesiohornblende
nserpentinegroupminerals(lizardite,antigorite)
or pargasite), anorthite (plagioclase, 5–20%), quartz (5–
Othermineralscomparedagainstthedata,butnotincludedinTable2
20%) and silvery retrograde margarite, with or with-
duetopoormatches:calcite,fluorite,corundum,topaz,anatase,rutile,
outaccessorygarnet(0–20%),rutile(0–3%),corundum,
variscite.
Geoarchaeology:AnInternationalJournal29(2014)121–137 Copyright(cid:2)C 2014WileyPeriodicals,Inc. 129
PROVENANCEOFLITTLESALTSPRINGGREENSTONES BONOMOETAL.
Figure8 (A)StereoscopicmicroscopeimageofarutilecrystalinA04.(B)SEM-EDSspectrumofarutilecrystalinA04,operatingwithanaccelerating
voltageof20.0kV.(C)StereoscopicmicroscopeimageofawhitemicaphaseinA04.(D)SEM-EDSspectrumofamonazitecrystalinA04,operatingwithan
acceleratingvoltageof20.0kV.(E)Clockwise,fromtop-left:secondaryelectronimageofamonazitecrystalinA03,with20μmscalebarintheupper-left;
PKαelementdistributionmap;NdLαelementdistributionmap;CeLαelementdistributionmap.
augite,spinel,kyanite,zoisite,orchromite(Pratt&Lewis, in other lithologies throughout the southeastern United
1905; Emilio, 1998; Emilio & Ryan, 1998; Peterson & States, for example, the Soapstone Ridge metapyroxen-
Ryan, 2009; Collins, 2011; Graybeal et al., 2012). Mon- ite/metaultramaficcomplexinGeorgia(Turner&Swan-
azite,thoughnotspecificallyidentifiedintheBuckCreek son, 1998; Chaumba, 2009), the Six Mile thrust sheet,
assemblages, is a common accessory mineral in some Walhalla nappe, and Chauga belt amphibolites in South
hydrothermally altered rocks (Zhu & O’Nions, 1999) Carolina (Prince & Ranson, 2004), granitoids of the Lib-
and is found in several belts throughout the southeast- erty Hill pluton in South Carolina (Speer, 1987), eclog-
ern Appalachian Piedmont (Mertie, 1979). Despite the ites of the eastern Blue Ridge of North Carolina (Page,
potentially varied mineralogy of the amphibolites, Pratt Essene, & Mukasa, 2004), and pyroxenites of the
and Lewis (1905) identify assemblages containing only Webster-Addie complex in North Carolina (Warner &
edenite and plagioclase. While A04 lacks any visible fo- Swanson, 2010), the edenite is typically subordinate to
liation suggestive of schist, green aluminous anorthite- other amphiboles (actinolite, magnesiohornblende, etc.)
bearing edenite amphibolites with accessory chromite, orotherwisepresentonlyinlowmodalabundance.
corundum, and spinel were previously identified in the The XRF analysis of A04 conducted at the Ion Beam
complexbyPrattandLewis(1905),andHadley(1949)as LaboratoryatSUNY—Albanyidentifiedtraceamountsof
alterationproductsofthetroctoliteandareequivalentto Rb, Sr, Zr, and Ba (Figure 6B); pXRF analysis corrobo-
theedenite-margariteschistsoflaterstudies(Peterson& rates the concentrations for Rb, Sr, and Zr. Relative to
Ryan, 2009); as such, the amphibolite-schist distinction A03, A04 lacks appreciable concentrations of Y and Nb.
maybeoflittleimportanceoutsideofastrictlygeological In addition to these trace elements, the pXRF spectrum
context.Whileedenite-bearingassemblagesareidentified also shows lower amounts of Ti and Fe relative to A03
130 Geoarchaeology:AnInternationalJournal29(2014)121–137 Copyright(cid:2)C 2014WileyPeriodicals,Inc.