Accessing Hunter-Gatherer site structures using Fourier transform infrared spectroscopy: applications at a Taltheilei settlement in the Canadian Sub-Arctic Don H. Butler * , Peter C. Dawson Department of Archaeology, University of Calgary, 2500 University Dr. N.W., Calgary, Alberta T2N 1N4, Canada article info Article history: Received 7 August 2012 Received in revised form 22 October 2012 Accepted 24 November 2012 Keywords: FTIR Hunter-gatherers Site structure Canada Hearths Bone burning Calcite Carbonate hydroxylapatite Authigenic phosphates Crystallinity index Carbonate/phosphate ratio abstract The results of Fourier transform infrared spectroscopy on soils and caribou bone from a Taltheilei culture settlement in northern Canada contribute to developing micro-archaeological approaches suitable for locating and characterizing hearth and midden features on hunter-gatherer sites. A weak yet pervasive signal for montgomeryite was developed from the diagenesis of dispersed ash and caribou processing residues. Disordered calcite, carbonate hydroxylapatite, charcoal, and burned bone in two pit-house hearth deposits indicate that both wood and bone were used for fuel. Crystallinity indices and carbonate/phosphate ratios for bone indicate high intensity burning. These data, in tandem with the presence of semi-subterranean dwellings, demonstrate that this particular tundra-based encampment was occupied during cold seasons, a type of settlement behaviour previously unrecognized in the Taltheilei archaeological record. Our results confirm that Fourier transform infrared spectroscopy is an accessible, rapid, and cost effective means of discovering micro-archaeological evidence valuable for reconstructing hunter-gatherer site structures. Ó 2012 Elsevier Ltd. All rights reserved. 1. Introduction 1.1. Objectives The archaeological record of Taltheilei caribou hunters (ca. 600 B.C.eA.D. 1700) is gradually being uncovered in southern Nunavut and northern Manitoba, Canada (Gordon, 1975, 1996; Nash, 1970; Petch, 1992; Petch et al., 1997). Taltheilei sites in the region are generally lithic scatters, stone markers, or small sites with low artefact densities that are interpreted as ephemeral camps (Schwimmer et al., 1998). Recent surveys and excavations in southern Nunavut have discovered new evidence documenting Taltheilei land-use, specifically the extended cold season use of tundra-based pit-house settlements (Dawson et al., 2009; Hodgetts et al., 2011). These sites provide unique opportunities to document the use and organization of space at the settlement scale. Locating and characterizing hearth and refuse disposal areas, in particular, are fundamental to accurately reconstructing space at hunter- gatherer settlements (Binford, 1983; Oswald, 1984). Centring on the micro-archaeological record at the Ikirahak settlement (JjKs-7) in southern Nunavut, this research uses Fourier transform infrared spectroscopic (FTIR) analyses of soils and caribou bone to charac- terize dwelling hearth features and to locate outdoor hearths and bone middens. We aim to determine the use of space at the site and clarify its season of occupation, which will contribute both to addressing issues with our understanding of Taltheilei settlemente subsistence patterns and to developing micro-archaeological approaches suitable for defining hunter-gatherer site structures. Frequencies, sizes, and locations of hearth and midden features are useful for reconstructing dynamics in site functions, seasons of occupation, mobility patterns, and socio-economics. The types of materials used for fuel can provide evidence for seasonality, while multiple overlapping hearths act as a record of reoccupation (Bamforth et al., 2005; Schiegl et al., 2003). Similarly, numerous large middens are common at long-term camp sites and at sites with long histories of reoccupation. Shorter term camps tend to develop a light layer of homogenously dispersed refuse (Kent, 1999). Smaller middens situated adjacent to dwellings typically represent household work and refuse disposal, while larger middens in central and peripheral areas can relate to communal activities (Beck and Hill, 2004; Oetelaar, 1993). * Corresponding author. Tel.: þ1 403 478 6739; fax: þ1 403 282 9567. E-mail address: [email protected](D.H. Butler). Contents lists available at SciVerse ScienceDirect Journal of Archaeological Science journal homepage: http://www.elsevier.com/locate/jas 0305-4403/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jas.2012.11.015 Journal of Archaeological Science 40 (2013) 1731e1742
Transcript
at SciVerse ScienceDirect
Journal of Archaeological Science 40 (2013) 1731e1742
Contents lists available
Journal of Archaeological Science
journal homepage: http : / /www.elsevier .com/locate/ jas
Accessing Hunter-Gatherer site structures using Fourier transform infraredspectroscopy: applications at a Taltheilei settlement in the Canadian Sub-Arctic
Don H. Butler*, Peter C. DawsonDepartment of Archaeology, University of Calgary, 2500 University Dr. N.W., Calgary, Alberta T2N 1N4, Canada
a r t i c l e i n f o
Article history:Received 7 August 2012Received in revised form22 October 2012Accepted 24 November 2012
Keywords:FTIRHunter-gatherersSite structureCanadaHearthsBone burningCalciteCarbonate hydroxylapatiteAuthigenic phosphatesCrystallinity indexCarbonate/phosphate ratio
0305-4403/$ e see front matter � 2012 Elsevier Ltd.http://dx.doi.org/10.1016/j.jas.2012.11.015
a b s t r a c t
The results of Fourier transform infrared spectroscopy on soils and caribou bone from a Taltheilei culturesettlement in northern Canada contribute to developing micro-archaeological approaches suitable forlocating and characterizing hearth and midden features on hunter-gatherer sites. A weak yet pervasivesignal for montgomeryite was developed from the diagenesis of dispersed ash and caribou processingresidues. Disordered calcite, carbonate hydroxylapatite, charcoal, and burned bone in two pit-househearth deposits indicate that both wood and bone were used for fuel. Crystallinity indices andcarbonate/phosphate ratios for bone indicate high intensity burning. These data, in tandem with thepresence of semi-subterranean dwellings, demonstrate that this particular tundra-based encampmentwas occupied during cold seasons, a type of settlement behaviour previously unrecognized in theTaltheilei archaeological record. Our results confirm that Fourier transform infrared spectroscopy is anaccessible, rapid, and cost effective means of discovering micro-archaeological evidence valuable forreconstructing hunter-gatherer site structures.
� 2012 Elsevier Ltd. All rights reserved.
1. Introduction
1.1. Objectives
The archaeological record of Taltheilei caribou hunters (ca.600 B.C.eA.D. 1700) is gradually being uncovered in southernNunavut and northern Manitoba, Canada (Gordon, 1975, 1996;Nash, 1970; Petch, 1992; Petch et al., 1997). Taltheilei sites in theregion are generally lithic scatters, stone markers, or small siteswith lowartefact densities that are interpreted as ephemeral camps(Schwimmer et al., 1998). Recent surveys and excavations insouthern Nunavut have discovered new evidence documentingTaltheilei land-use, specifically the extended cold season use oftundra-based pit-house settlements (Dawson et al., 2009; Hodgettset al., 2011). These sites provide unique opportunities to documentthe use and organization of space at the settlement scale. Locatingand characterizing hearth and refuse disposal areas, in particular,are fundamental to accurately reconstructing space at hunter-gatherer settlements (Binford, 1983; Oswald, 1984). Centring on
: þ1 403 282 9567.).
All rights reserved.
the micro-archaeological record at the Ikirahak settlement (JjKs-7)in southern Nunavut, this research uses Fourier transform infraredspectroscopic (FTIR) analyses of soils and caribou bone to charac-terize dwelling hearth features and to locate outdoor hearths andbonemiddens. We aim to determine the use of space at the site andclarify its season of occupation, which will contribute both toaddressing issues with our understanding of Taltheilei settlementesubsistence patterns and to developing micro-archaeologicalapproaches suitable for defining hunter-gatherer site structures.
Frequencies, sizes, and locations of hearth and midden featuresare useful for reconstructing dynamics in site functions, seasons ofoccupation, mobility patterns, and socio-economics. The types ofmaterials used for fuel can provide evidence for seasonality, whilemultiple overlapping hearths act as a record of reoccupation(Bamforth et al., 2005; Schiegl et al., 2003). Similarly, numerouslarge middens are common at long-term camp sites and at siteswith long histories of reoccupation. Shorter term camps tend todevelop a light layer of homogenously dispersed refuse (Kent,1999). Smaller middens situated adjacent to dwellings typicallyrepresent household work and refuse disposal, while largermiddens in central and peripheral areas can relate to communalactivities (Beck and Hill, 2004; Oetelaar, 1993).
Hearths and middens are often faintly expressed or eveninvisible in the archaeological record, and their locations may onlybe defined through micro-archaeological techniques such as soil/sediment chemistry (Middleton and Price, 1996). Fourier transforminfrared spectroscopy of soils, sediments, and reference materials isan accessible, rapid, and cost-effective means of locating fadinghearth and refuse disposal features on hunter-gatherer sites. Herewe use FTIR to define the impact of burning and bone decay on themineral composition of soils from Ikirahak’s dwelling hearths andoutdoor spaces and to determine the temperatures bones exca-vated from hearths were exposed to during burning. The influenceof burning on soils is investigated using traces of calcite, carbonatehydroxylapatite, and hematite, while the effects of bone decom-position are explored using the presence of authigenic phosphates(Karkanas et al., 2007; Weiner et al., 1993). Bone burning is eval-uated using the crystallinity index, the carbonate/phosphate ratio,and changes in organic matter, collagen, and phosphate peakcharacteristics (Squires et al., 2011; Thompson et al., 2009). Ourresearch is among the first using FTIR to characterize materialsfrom an archaeological site in the Canadian Arctic (e.g., Helwiget al., 2008) and it provides preliminary evidence for the forma-tion of authigenic montgomeryite on an open-air archaeologicalsite.
1.2. Taltheilei caribou hunters
The Taltheilei archaeological culture was defined throughexcavations at several sites around the Taltheilei Narrows of GreatSlave Lake (Nash,1970). Its roots originate in the forests of the PeaceRiver Valley, northern British Columbia and Alberta (Gordon, 1975).Between 700 and 600 B.C. pioneers moved northeast into theBeverly Caribou Range. Radiocarbon dates from forest sites on theDubawnt River demonstrate that the earliest Taltheilei groupsrapidly populated the area and adapted to hunting caribou around575 B.C. (Gordon, 1996) (Fig. 1). Gordon (1996) argues thatthroughout their roughly 2300 year occupation of the region, they
Fig. 1. The central Can
spent spring and fall following caribou to and from the tundrarespectively. During midsummer, they lived on the tundra in small,highly mobile groups. Winters were passed in the forest in larger,more sedentary groups.
Over 1000 Taltheilei sites (ca. 600 B.C.eA.D. 1700) have beendiscovered throughout Beverly Caribou Range’s tundra and borealforest landscapes (Gordon, 1996). Far less research has beenundertaken in the Kaminuriak Caribou Range, though recentarchaeological surveys around Maguse Lake identified severalpit-house sites (Dawson et al., 2009; Hodgetts et al., 2011) (Fig. 1).These are among the first identified Taltheilei sites of their kind.The Ikirahak site (JjKs-7), the focus of this paper, is a tundra-basedpit-house settlement located on the southwestern shoreline ofa small island that sits in the Narrows of Maguse Lake near a largecaribou crossing. Positioned on a gently sloping terrace approxi-mately 10 m above the lake shore, the site consists of 10 pit-housedepressions and several potential storage features (Fig. 2). Diag-nostic points and hide abraders, living floors, and AMS dates oncaribou bone indicate that Taltheilei people occupied the settle-ment several times between approximately A.D. 500 and 650.
An investment into building pit-houses supports the argumentthat this tundra-based settlement was occupied for a longer periodof time during a cold season (Binford, 1990; Kelly et al., 2005;LeMouel and LeMouel, 2001; McGuire and Schiffer, 1983). Thisdiffers significantly from the settlement pattern documented in theBeverly Caribou Range, where the tundra was occupied by smallmobile groups during the summer. The houses are each approxi-mately 4.5 m in diameter, and they were constructed by excavatinga roughly 50 cm deep circular pit and banking the sedimentsaround the edge. Boulders were placed around the perimeter of thesediment berms to secure hide superstructures. The rocks werepushed into the centres of the dwellings upon abandonment. Agroup of caribou hunters occupying the site for an extended periodwould produce a substantial amount of refuse, yet there are nodiscernible bone middens at Ikirahak. Bone may have decomposedand reprecipitated in soils as authigenic phosphates and/or it was
used as fuel, both of which are evaluated below using FTIR. Anal-yses of burned bone will also help clarify whether the site wasoccupied during a cold season.
2. Locating hearth and bone midden features using FTIR
2.1. FTIR
Fourier transform infrared spectroscopy measures the vibra-tional behaviours of chemical bonds (Bhargava and Levin, 2005).Infrared (IR) light passing through a molecule causes its bonds tovibrate in characteristic ways, such as asymmetric stretching.Vibrations are recorded as frequencies that are Fourier transformedinto wavenumbers (the inverse of wavelength) (Smith, 2011).Specific types of bonds absorb characteristic amounts of infraredradiation at diagnostic wavenumbers between 250 and 4000 cm�1
(Harwood and Claridge, 1997). A characteristic carboneoxygenbond in carbonate, for instance, exhibits an asymmetric stretchbetween 1415 and 1420 cm�1 (Tatzber et al., 2007).
The strengths and applications of FTIR are far-reaching andinclude identifying functional groups of molecules, defining atomicstructures, and characterizing unknown materials. Straightforwardsample preparation and instrument operation procedures makeFTIR analyses of systematically collected soil/sediment samples anaccessible and rapid means of characterizing and locating a varietyof feature types. Three-dimensional spatial analyses of mineralassemblages are particularly valuable in reconstructions of bothsite formation processes and the structured use of space (Weineret al., 2002). Instruments measure a range of wavenumberssimultaneously and they rapidly take multiple measurements ofsamples, providing averaged spectra with optimal signal-to-noiseratios (Bhargava and Levin, 2005). In addition, instrumentsproduce accurate and replicable results in laboratory and fieldcontexts (Shahack-Gross et al., 2005), they simultaneously measureorganic and inorganic compounds (Thompson et al., 2009), andthey identify atomically disordered minerals such as calcite derivedfrom wood ash, authigenic phosphates from decomposing ash andbone, and carbonate hydroxylapatite in bone (Karkanas et al., 2000;
Weiner, 2010). Drawbacks include intra-sample variation owing totiny aliquot sizes, shifting peak maxima owing to variation ingrinding, and overlapping absorbance peaks in mixtures such assoils (Smith, 2011).
2.2. Hearth features
The clay, iron, and calcite components of soils and sediments,defined using IR spectra, are useful for distinguishing the locationsof primary hearth features and secondary ash dumps when visiblematerial evidence is faded or erased. Intense heat causes disorder inthe atomic structures of clay minerals in adjacent soils and sedi-ments. Montmorillonite, for instance, undergoes several diagnosticstructural changes between 400 and 1100 �C. Disorder in themineral, indicated by shifting peaks, the disappearance of peaks,and the appearance of sharp peaks, is related to the temperature ofthe heat source, distance from the heat source, and the duration ofexposure (Berna et al., 2007). Rubification owing to high concen-trations of hematite also helps distinguish primary hearth contexts.Acidic and reducing conditions caused by burning increase thesolubility of iron in the deposit. Reduced iron tends to dissolve andre-oxidize into the insoluble iron mineral hematite, which bothreddens the deposit and enhances its magnetic signal (Aspinallet al., 2008; Karkanas et al., 2002). Heat altered hematite hasseveral sharp IR absorbance bands between 650 and 400 cm�1
(Rendon and Serna, 1981).A more frequently applied approach to defining hearths using
FTIR focuses on disordered calcite in soils and sediments (Albertet al., 2003; Goldberg et al., 2012; Karkanas et al., 2002; Schieglet al., 2003; Weiner et al., 1995, 1998, 2002). Burning wood attemperatures between 400 and 500 �C transforms its calciumoxalate component into calcite through the expulsion of carbonmonoxide (Huaqing, 1989; Regev et al., 2010). Calcite absorbs at1420 cm�1 (n3, asymmetric stretch), 874 cm�1 (n2, out-of-planebending), and 713 cm�1 (n4, in-plane bending) (Chu et al., 2008).The ratio of the n2 and n4 peaks distinguishes the mineral’s level ofatomic disorder. Ash calcite is disordered, while the geogenicpolymorph is much more crystalline (Regev et al., 2010).
Pyrogenic calcite, however, may not survive site formingprocesses (Karkanas et al., 2007). It is relatively soluble andunstable, specifically in the acidic, organically enriched conditionscaused by burning. In acidic deposits rich in phosphates, amor-phous calcite can be transformed into the more stable authigeniccarbonate hydroxylapatite (Weiner et al., 1993). The formation ofsuch authigenic phosphates in soils and sediments is primarilylinked to acidity, phosphate levels in the system’s water, mineralstability, and available carbonates (Karkanas et al., 2000; Nriagu,1976). Decomposing human and animal wastes, vegetation, bone,and many other types of organic matter add significant amounts ofphosphate to soils and sediments, which in turn causes a reductionin the pH of the deposit (Goldberg and Nathan, 1975). Theseconditions contribute to the diagenesis and reprecipitation ofcarbonates, including amorphous ash calcite and carbonatehydroxylapatite from decaying bones, into authigenic phosphates(Karkanas et al., 2002).
2.3. Bone as a fuel source
Hearths containing few wood charcoal fragments and minorchemical traces of calcite, but abundant calcined bone fragmentsand strong signatures for primary carbonate hydroxylapatite,indicate that both bone and wood were burned, yet bone was theprincipal fuel source (Schiegl et al., 2003). Stages of bone burningand natural decomposition can be identified using IR spectra todefine the crystallinity index (CI, also called the splitting factor), the
carbonate/phosphate ratio (C/P), and organic matter, collagen, andphosphate peak changes (Thompson et al., 2009).
Several recent studies detail the strengths and weaknesses ofthe CI for defining the characteristics of burned bone from botharchaeological and experimental contexts (Lebon et al., 2010;Squires et al., 2011; Thompson et al., 2009, 2010). The degree ofsplitting between the peaks of the carbonate hydroxylapatitedoublet at 605 and 565 cm�1 corresponds to the organization ofbone tissue’s atomic structure (Nagy et al., 2008; Weiner and Bar-Yosef, 1990). As natural decomposition or burning progresses, theatomic structure of carbonate hydroxylapatite becomes moreorganized (Stiner et al., 1995; Weiner and Price, 1986). Mineralcrystals become larger, more crystalline, and more chemicallystable (Shipman et al., 1984). Sharper absorption peaks at 605 and565 cm�1 indicate an ordered crystalline structure (Thompsonet al., 2009). Changes in peak sharpness are a function of atomicalterations caused by burning, but they are also caused by theslower processes of chemical diagenesis (Stiner et al., 2001). Assuch, the crystallinity index alone may not be an accurate indicatorof bone burning (Lebon et al., 2010; Trueman et al., 2008). Here, it issupplemented with the C/P ratio and changes in organic matter,collagen, and phosphate peak characteristics (Weiner, 2010).
The C/P ratio provides an estimate of the amount of carbonate ina sample (Thompson et al., 2009). It decreases as the carbonatefraction decreases (Squires et al., 2011). The phosphate componentused in the ratio absorbs between 1035 and 1048 cm�1, while thecarbonate component absorbs between 1415 and 1420 cm�1. Bothhighly burned and decomposed bone contain close to no carbonate,indicated by an absence of peaks at 872, 1420, and 1456 cm�1
(Stiner et al., 1995). Bones burned at high temperatures, typicallyexceeding 700 �C, also have sharper, narrower phosphate peaks(Thompson et al., 2009). Moreover, the phosphate peak shifts from1035 cm�1 in unaltered bone to 1048 cm�1 in highly burned bone(Weiner, 2010). The organic matter doublet at 2924/2853 cm�1 andcollagen peaks at 1750e1550 cm�1 and 1455 cm�1, representingamide I and carbonyl functional groups respectively, are drasticallyreduced in highly mineralized/burned bone. Collagen peaks areprimarily absent in samples that have been heated above 700 �C(Thompson et al., 2009). The appearance of sharp peaks at roughly632 and 1090 cm�1 represent the formation of crystalline phos-phate and apatite phases, making them some of the best indicatorsof calcined bone (Weiner, 2010).
2.4. Bone diagenesis after burial
Infrared spectra of soils and sediments can help pin-point loca-tions where bone was deposited and subsequently decomposed, orwhere ash calcite has undergone significant diagenesis and repre-cipitation (Berna et al., 2004; Weiner et al., 1993). Bone collagendecomposes in alkaline soil/sediment environments, leavinga carbonate hydroxylapatitemineral ghostwith lowporosity (Collinset al., 2002; Nielsen-Marsh and Hedges, 1999). The carbonatehydroxylapatite component decomposes in acidic, aerobic deposits(Cronyn, 1990; Hedges, 2002). Under these circumstances, bonegradually dissolves and its carbonate mineral products can reactwith available phosphates to form relatively insoluble authigenicphosphates (Karkanas et al., 1999; Schiegl et al., 1996).
A reaction cascade for authigenic phosphates in archaeologicalcave sediments has been defined and used to establish where boneswere deposited, completely decayed, and reprecipitated as newmineral formations, to distinguish primary from secondary deposits,and to distinguish episodes of occupation and abandonment(Karkanas et al., 2002; Schiegl et al., 1996; Weiner et al., 1993).Crandallite forms in acidic deposits (pHw 4e6)with high aluminumandcalciumyet lowphosphate levels,whilemontgomeryite forms in
acidic deposits (pH w 4e6) with high phosphate, calcium,aluminum, and magnesium levels (Karkanas et al., 2000; Nriagu,1976; Weiner, 2010). Variscite and taranakite can form after pro-longed alteration of montgomeryite or crandallite in highly acidicdeposits (pH< 4)with elevatedphosphate concentrations (Goldbergand Nathan, 1975; Karkanas et al., 2002). Taranakite can also formdirectly from carbonated hydroxylapatite in neutral pH environ-ments (Karkanas et al., 2000). Authigenic phosphates reprecipitatedfrom calcite or carbonate hydroxylapatite can be atomically disor-dered. Peaks for the disordered montgomeryite polymorph formingfrom these minerals characteristically appear at 1070, 1058, and595 cm�1, and also at approximately 3413, 1643, 1036, 590, and562 cm�1 (Weiner et al., 2002; Weiner, 2010).
3. Materials and methods
3.1. Sampling and processing
One hundred forty-five soil samples were collected from thesettlement on a lattice-grid using an Oakfield soil corer. A 3 msample interval was chosen to balance the size of the samplinguniverse (w5 km2), the potential sizes of areas influenced byoutdoor hearths and bone middens, and time in the field (Fig. 2).Some gaps in our grid exist because samples could not be takenfrom every point. Some areas were too rocky, had poor soilformation, were disturbed by excessive cryoturbation, or over-grown with brush. During excavations, samples were collectedfrom the hearths of Houses 3 and 8 and from a potential hearth inHouse 2. Addressing intra-sample variation, we first homogenizedour large bulk samples using a porcelain pestle and mortar. Weseparated the fine fractions using a 2 mm sieve and then micro-chute split (riffling) them to reduce the sample size. Aliquotswere taken from the reduced samples and oven dried at 120 �C for24 h to remove gravimetric water.
Organic matter contents were determined using the loss-on-ignition method and particle size analyses were conducted usingaMalvern laser diffraction system. The pH and Ehwere recorded foreach sample in a 1:2 mixture of soil and distilled water usingpotentiometric soil meters. Element concentrations weremeasuredas the percent weight of their oxides using x-ray fluorescence (seeButler, 2011).
Archaeological caribou bone samples (Rangifer tarandus groen-landicus) were collected during the excavations of Houses 2, 3, and 8(floorN¼16; hearthN¼6). Fresh (N¼ 4), burned (N¼ 3), andburied(N ¼ 1) comparative samples were collected from two recentcaribouprocessing campsnear thehamlet of Arviat. To reduce issuescaused by intra-sample variation, we selected relatively largesamples for crushing. Samples were crushed using a Carver Labo-ratory press. We homogenized the particles using a porcelain pestleand mortar and we then separated a very fine fraction usinga 0.15 mm sieve. This fraction was micro-chute split and an aliquotwas taken from the reduced sample. Samples were also character-ized using preservation categories described byHaynes et al. (2002).
3.2. FTIR analyses
Absorption spectra were collected using the potassium bromide(KBr) disk method (Surovell and Stiner, 2001). Approximately300 mg of KBr powder was gently ground and homogenized with2e4 mg of sample. Although grinding was facilitated manuallyusing a porcelain pestle and mortar, we aimed to reduce the effectsof shifting peak maxima owing to variation in grinding by prac-ticing and duplicating the appropriate grinding force, pattern, andtiming. Roughly half the mixturewas evenly spread inside a KBr dieand pressed at 13,000 lbs on a Carver Press under a vacuum for 60 s.
Spectra were collected using a Nicolet Nexus 470 operated byOMNIC� 6.2 software. A background spectrumwas collected every100 min and a nitrogen gas flowwas used to eliminate interferencefrom atmospheric water and carbon dioxide. Spectra were derivedfrom an average of 64 scans at 4 cm�1 resolution, they were linearbaseline corrected, and they were smoothed using 13 points.
3.3. Hearth and midden features
Traces of amorphous calcite, primary carbonate hydroxylapatite,and hematite were used to search for outdoor hearth areas. Laserdiffraction particle size analyses did not detect any clay in thesamples, excluding the use of clay minerals to identify hearthlocations. Calcite was defined in the soil samples using its charac-teristic peaks at 1420 cm�1 (n3), 874 cm�1 (n2), and 713 cm�1 (n4).Following Chu et al. (2008), we used the ratio of the n2 and n4 peakheights to determine whether any discovered calcite is wood ash orgeological. Measuring baselines were drawn from the lowest pointsof the valleys bookending the peaks. Geological calcite has valuesaround 3.0, while values for ash calcite are around 4.0. The presenceof heat altered hematite was assessed using sharp peaks at 650,595, 525, 470, 440, and 400 cm�1 (Rendon and Serna, 1981).
We explored the site for diagenetically transformed bonemiddens using authigenic phosphates. To identify these minerals,we compared our spectra with literature data (Karkanas et al.,2002; Schiegl et al., 1996; Weiner et al., 2002; Weiner, 2010) andwith archived reference spectra (KCAS, n.d.) for montgomeryite,crandallite, variscite, and taranakite. Areas that have no evidence ofbone fragments, carbonate hydroxylapatite, or authigenic phos-phates are considered free of alteration caused by the deposition,diagenesis, and reprecipitation of bone mineral. An absence ofcalcite and carbonate hydroxylapatite along with the presence ofdisordered montgomeryite will indicate the diagenesis and trans-formation of phosphates and carbonates in an acidic environment.On archaeological sites, refuse from human activities, specificallybone and ash particulates, provides abundant phosphates,carbonates, and magnesium that may contribute to the formationof authigenic phosphates (Goldberg and Nathan, 1975). The pres-ence of authigenic montgomeryite in an archaeological context isa dependable indicator that bone and/or ash were deposited butdid not preserve (Karkanas et al., 2007; Weiner et al., 1993).
3.4. Bone crystallinity
Crystallinity indices were calculated by dividing the sum of the605 and 565 cm�1 peak heights by the height of the valley sepa-rating them (Weiner and Bar-Yosef, 1990). Large values indicatesignificant decomposition and structural ordering of the boneapatite (Lebon et al., 2010; Munro et al., 2007). Values for modernunaltered bone typically lie between 2.5 and 3.5. Values formoderately decomposed or burned bone range between approxi-mately 3.6 and 4.5, while those for highly decomposed or burnedbone range between 5 and 7 (Thompson et al., 2009).
Carbonate/phosphate ratios were calculated by dividing theheight of the carbonate peak at 1420 cm�1 with the height of thephosphate peak at 1035 cm�1 (Squires et al., 2011). Unaltered bonehas a ratio between 0.31 and 0.65. Decreasing values are a functionof natural diagenesis, but they are exacerbated by burning. Valuesas low as 0.04 indicate burning at temperatures surpassing 700 �C(Thompson et al., 2009). Supplementing the CI and C/P ratio,degrees of burning/mineralization were classified using changes inorganic matter, collagen, and phosphate peaks. We use thedecomposition of the organic matter doublet at 2924/2853 cm�1
and collagen peaks at 1750e1550 cm�1 and 1455 cm�1, along withsharper, narrower, left-shifted phosphate peaks to identify highly
mineralized/burned bone. Additionally, the formation of sharppeaks on the carbonate hydroxylapatite doublet at 632 cm�1 and onthe phosphate peak at 1090 cm�1 will indicate the formation ofcrystalline mineral phases (Weiner, 2010).
Descriptive statistics and analysis of variance (ANOVA) under-taken using S-Plus 8.0 and SPSS 16.0 were used to distinguishgroups of unaltered, buried, and burned bone. The analysis tests thenull hypothesis that there is no difference in CI and C/P valuesbetween the groups using the F-test to compare the variances ofmeans within groups and between groups. To satisfy statisticalassumptions and avoid type I statistical error, quantileequantileplots were used to assess whether the sample population is nor-mally distributed and Levene’s score was used to test for homo-geneity of variances.
4. Results
4.1. Soil analyses
Soils at the site are organically enriched cryogenic silt loams.Loss-on-ignition organic matter content ranges from 16% to 83%(x
¼ 51%). Laser diffraction particle size analyses indicate the soilscontain an average of 42% sand and 58% silt. They are also acidic andaerobic, having pH readings between 3.76 and 5.00 (x
¼ 4.22) andoxidizing redox potentials between 408.30 mv and 519.50 mv(x ¼ 467.81 mv) (Table 1). Cryoturbation is indicated by largesurface features such as ice wedges and hummocks and micro-morphologically by icewedging andplanar voids in soil thin sections(McNamee et al., 2009). Charcoal is present in the House 3 and 8hearth samples and in several cores from across the site. Burnedbone fragments are present in the House 3 and 8 hearth samples.
All of the collected spectra have phosphate and carbonatecomponents. Calcite and carbonate hydroxylapatite are present inthe hearth soils of Houses 3 and 8. Calcite is disordered, with n2/n4ratios of 4.5 for House 3 and 4.0 for House 8. Calcite, carbonatehydroxylapatite, and hematite are not present in the soils fromacross the site. Montgomeryite is near ubiquitous across the site,but is absent in the dwelling soils (Fig. 3). Some of its peaksare small, indicating weak absorbencies, and in turn, weakconcentrations. Only 25 of the 148 samples lack peaks for themineral. X-ray fluorescence also identified enriched concentrationsof phosphorus, calcium, aluminum, and magnesium in the sitesamples relative to those collected away from the site (Table 1).
4.2. Bone analyses
Median, minimum, and maximum CI and C/P values areprimarily distinct for the three groups, but their ranges have some
overlap (Fig. 5; Table 2). Means for each group are also somewhatdifferent but there are some issues with large standard deviations,which could be reduced by using a larger sample size. Crystallinityindex means for unaltered, buried, and burned bones were 2.6, 3.2,and 4.2 respectively. The standard deviation of the unaltered bone(s ¼ 0.08) is much lower than those for the buried (s ¼ 0.60) andburned (s ¼ 0.88) bones. The highest value for CI, 5.5, is froma burned bone retrieved from the hearth in House 3. The lowestvalue for C/P, 0.06, is from the same sample. The C/P means are0.62, 0.19, and 0.13 for unaltered, buried, and burned groups.Standard deviations are somewhat large, valuing 0.26 and 0.12 forunaltered and buried bone respectively. Burned bone was muchlower at 0.06.
Analysis of variance results, supporting the descriptive statistics,highlight statistically significant differences in the CI and C/Pmeansbetween unaltered, buried, and burned bone groups. Quantileequantile plots display normal distributions for both CI and C/Pand values for Levene’s test (p � 0.05) indicate homogeneousvariances, both contributing to the robusticity of the analysisagainst type I error. F-values for CI and C/P, 13.04 and 22.43respectively, are statistically significant and they far-exceed thecritical Fevalue (F ¼ 5.49; df ¼ 2, 27; p � 0.01), indicating thegroups are statistically distinct. Tukey post-hoc tests for CI confirmthat unaltered and burned bone have the greatest differences.Buried and burned groups are also significantly different. Unalteredand buried groups have some overlap. Means for the C/P ratiodiverge considerably between the unaltered and buried groups andbetween the unaltered and burned groups. Burned and buriedgroups have some overlap caused by their standard deviations.Based on the ANOVA and descriptive statistics, the CI distinguisheschemically decomposed bone (buried) from burned bone betterthan the C/P ratio in our dataset (Fig. 5).
We also documented clear changes in organic matter, collagen,and phosphate peaks (Fig. 4). The organic matter doublet is veryhigh in fresh bone, yet significantly reduced in the burned/decomposed samples. Collagen peaks are extensively reduced inheight and width during the sequence of decomposition. Sampleswith high CI and low C/P values have sharper phosphate peaksthat are shifted slightly from the 1030 s to 1040 s cm�1. Threehighly resolved, sharp peaks also develop throughout decompo-sition process, and they distinguish highly mineralized/burnedbone samples. Sharp peaks appear on the left side of the phos-phate peak at 962 cm�1 and on the right side of the phosphatepeak at 1090 cm�1, representing the formation of a highly crys-talline phosphate phase. Similarly, a highly resolved peak formson the left side of the carbonated hydroxylapatite doublet at632 cm�1, which again relates to the formation of a crystallinephase.
5.1. Accessing hunter-gatherer site structures using FTIR
Fourier transform infrared spectroscopy is an accessible, rapid,and cost-effective means of collecting micro-archaeologicalevidence useful for uncovering the site structures of hunter-gatherer settlements. Instruments are widely available andsample processing and analyses are comparatively fast andstraightforward. We processed and analysed on average 7 samples
an hour and we logged 25 h, making our project both efficient andinexpensive. Considering the rapidity and low cost, soil surveysusing FTIR assist the identification of areas on sites for further, moreadvanced and costly analyses such as inductively coupled plasmamass-spectroscopy and gas chromatography.
Fourier transform infrared spectroscopic analyses contribute toidentifying the characteristics and locations of features on archae-ological sites that may otherwise remain invisible. Hunter-gatherers engaged in many types of spatially organized work thatdid not leave behind common types of material evidence. Many
tasks did, however, introduce chemical residues into underlyingsoils. Facilities that may remain invisible to routine investigationyet leave chemical footprints include storage areas, hide processingareas, meat drying areas, hearth and food preparation areas, refusedisposal areas, bedded areas, and houses built from snow and ice(Butler, 2011; Knudson and Frink, 2010; Middleton and Price, 1996).These footprints, in many cases, are required to accurately recon-struct the range of activities represented at settlements. Infraredspectra of soils, sediments, and reference materials provide datauseful for distinguishing the footprints of various types of high and
low diversity work spaces. As mentioned above, however, soils arecomplex mixtures of compounds, many of which absorb in similarareas of the IR spectrum. Overlapping absorbance peaks aresometimes difficult to identify, and they can, along with poorquality spectra, lead to the misidentification of the compoundspresent in a complex mixture like soil.
The chemical record at Ikirahak is clear in our IR spectra, whichowes to the high quality of our KBr discs. Disks were primarily clearand they produced spectra having absorbencies below one, yet notbelow zero, indicating the proper ratio of KBr and sample was used.
Fig. 5. Comparison of CI and C/P across unaltered, buried, and burned bone groups.
Opacity in some disks was recognized by the sloping background oftheir spectra, yet these slopes were not severe and were easilybaseline corrected. Spectra also lacked significant noise andcorrugation in the fingerprint region (1500e600 cm�1), indicatinga high signal-to-noise ratio and non-reflective disks (Weiner, 2010).It is not uncommon to use a low number of scans to increase pro-cessing and analysis efficiency, though this can sometimes diminishthe quality of the resulting spectra. Increasing the number of scansfor our analyses did not disrupt the efficiency of our processing andanalysis protocol, and, more importantly, it produced high qualityspectra that required only moderate post-processing.
Diagnostic peaks for calcite and those used to classify degrees ofbone decomposition/burning are well documented in the litera-ture, and they are clearly resolved in our spectra (Chu et al., 2008;Squires et al., 2011; Weiner and Bar-Yosef, 1990). We are alsoconfident that absorptions in our samples at 3413, 2348,1643,1384,
Table 2Bone characteristics.
Sample Context Description
1 Arviat; Recent Processing Camp 2 Surface Fresh; robu2 Arviat; Recent Processing Camp 2 Surface Fresh; robu3 Arviat; Recent Processing Camp 2 Surface Fresh; robu4 Arviat; Recent Processing Camp 2 Surface Sun bleach5 Arviat; Recent Cache; Buried Fragile edg6 Ikirahak; House 3; Floor; Buried Robust edg7 Ikirahak; House 3; Floor; Buried Robust edg8 Ikirahak; House 3; Floor; Buried Robust edg9 Ikirahak; House 3; Floor; Buried Robust edg10 Ikirahak; House 3; Floor; Buried Fragile edg11 Ikirahak; House 3; Floor; Buried Robust edg12 Ikirahak; House 2; Floor; Buried Robust edg13 Ikirahak; House 2; Floor; Buried Fragile edg14 Ikirahak; House 2; Floor; Buried Fragile edg15 Ikirahak; House 8; Floor; Buried Robust edg16 Ikirahak; House 8; Floor; Buried Robust edg17 Ikirahak; House 8; Floor; Buried Fragile edg18 Ikirahak; House 8; Floor; Buried Fragile edg19 Arviat; Recent Processing Camp 1 Fire Pit Burned; w20 Arviat; Recent Processing Camp 1 Fire Pit Burned; bl21 Arviat; Recent Processing Camp 2 Fire Pit Burned; w22 Ikirahak; House 8; Floor; Buried Burned; bl23 Ikirahak; House 8; Floor; Buried Burned; bl24 Ikirahak; House 8; Floor; Buried Burned; bl25 Ikirahak; House 8; Hearth; Buried Burned; bl26 Ikirahak; House 8; Hearth; Buried Burned; bl27 Ikirahak; House 8; Hearth; Buried Burned; w28 Ikirahak; House 3; Hearth; Buried Burned; w29 Ikirahak; House 3; Hearth; Buried Burned; w30 Ikirahak; House 3; Hearth; Buried Burned; w
representmontgomeryite. Our spectra share many similarities withthose reported by Weiner et al. (2002) and Weiner (2010). Peaks at1080e1070, 1058, and 595 cm�1 are particularly diagnostic ofmontgomeryite (Weiner et al., 2002). Defining the mineralassemblage of the soils also helps clarify any potential ambiguity.Silicates are common in soils, yet do not absorb in the same loca-tions as montgomeryite. Iron oxides, also very common in soils,cause broad, strong absorptions below 700 cm�1. These peaks arenot dominating this region our spectra. Clay minerals absorb insimilar wavenumber ranges as montgomeryite, though, in ourstudy, it is not possible that absorptions from clay minerals areleading to the misidentification of montgomeryite because particlesize analyses show there has been no clay formation, owing toreduced chemical weathering caused by cold temperatures. Humicacid is common in our spectra, and there is some potential overlapwith montgomeryite absorptions. Despite this, peaks for bothcompounds are generally distinguishable. In Sample 8, for example,the peak for humic acid at 475 cm�1 appears as a shoulder on theleft side of the 455 cm�1 montgomeryite peak, while the 535 cm�1
humic acid peak appears as a shoulder on the right side of the562 cm�1 montgomeryite peak (Fig. 3). X-ray fluorescence alsoconfirms the soils contain abundant aluminum, phosphorus,calcium, and magnesium, which are necessary for the formation ofmontgomeryite.
5.2. The structure of cold season stays at Ikirahak
Based on the results presented above, we propose that spatiallyconstrained bone middens were never formed at the site, thata substantial amountof bone refusewasused as fuel, that hearthfireswere hot enough to calcine bone, and that hearths were restricted todwelling interiors. The acidity, mineral components, and elementcontents of the soils provide an environment suitable for theformation of authigenic phosphates, particularly montgomeryite
(Table 1; Fig. 3). An absence of calcite and carbonate hydroxylapatitealong with the presence of montgomeryite throughout the siteprovides support for the diagenesis of deposited ash and/or boneparticulates (Weiner et al., 2002).
Montgomeryite is widespread across the site, making it unlikelythat it formed as a result of decomposing bone discarded inspatially discrete midden areas. There are no spatially patternedtraces of bone fragments or carbonate hydroxylapatite to supportspatially organized bone deposition. Moreover, the soils are frozenfor most of the year, and they are moderately acidic and oxidizing,meaning large fragments of buried bone would survive in thiscontext (Table 1). It is also unlikely that montgomeryite formednaturally through geological processes. Montgomeryite can form incalcareous, clay rich soils associated with phosphatic sedimentaryrocks (Nriagu,1976), yet, the deposit at Ikirahak would not facilitatethe development of montgomeryite in this way because it lacksgeogenic carbonate (no visible nodules or reactionwith dilute HCL),it lacks clay, and local parent geology primarily consists of granitictills and mafic igneous bedrock (Hodgetts et al., 2011).
An alternative explanation is that montgomeryite is a product ofthe diagenesis of dispersed ash and caribou processing residuescaused by the organically enriched, acidic, and aerobic soils acrossthe site (Table 1; Fig. 3). Ash, flesh, bone, skin, and hair particulates,bodily fluids, and tanning agents contributed the carbonates,phosphates, and magnesium necessary to form a weak yet perva-sive signal for montgomeryite. Traces of charcoal in several coresamples support that particulates were dispersed and depositedacross the site. Moreover, x-ray fluorescence results show generallyenhanced concentrations of phosphorus, aluminum, magnesium,and calcium across the site. As mentioned, these elements arerequired for the formation of montgomeryite. More importantly,enrichments in their concentrations have beenwidely documentedin association with the deposition of ash, charcoal, and liquid andparticulate residues from animal processing (Knudson and Frink,2010; Middleton and Price, 1996). Ash, charcoal, and decompos-ing organic matter from caribou hide and meat processing loweredthe pH of the deposit and provided an abundance of mobilephosphorus, calcium, and magnesium that interacted with geo-genic aluminum to start a reaction cascade culminating in theformation of montgomeryite.
Montgomeryite is absent in the dwelling samples, yet unstabledisordered calcite is present, demonstrating their soils are morechemically stable and less altered than those from the rest of thesite. Soils inside the dwellings contain more organic matter andwater, they are deeper, and they have good vegetation cover,making them colder for longer periods, interrupting mineraldiagenesis, and protecting them from the weathering processesaffecting shallower deposits throughout the site (Karkanas et al.,1999). Caribou processing residues and ash calcite dispersed overthe site experienced a greater degree of weathering and diagenesisthan the amorphous calcite and bone derived carbonate hydrox-ylapatite in the colder, unexposed hearth deposits. Althoughevidence for authigenic phosphates in various soils is available,additional research using experimental deposits, micromor-phology, electron microprobe, and lysimetry will clarify theprocesses involved with their formation and behaviour on open-airarchaeological sites. Distinguishing those formed from ash andbone is particularly important.
Several formationprocesseswould have contributed to removingbone refuse from the site. Soil forming processes and sedimentationrates are rather slow at Ikirahak, so it is unlikely that large bonefragments were buried. Furthermore, dogs were commonly kept byhistoric period Chipewyan people (Hearne, 1958). If Taltheileipeople, the ancestors of the Chipewyan, also kept dogs, they wouldhave fed them food refuse. Similarly, after abandonment some
remains would have been removed by scavenging animals. Peoplemay have also dumped refuse in the lake, or near the lake edge,where it was subsequently removed by ice rafting.
In addition, our results confirm the use of bone, and wood, tofuel dwelling hearths. People likely built hearths using drift woodand dwarf willow, adding bone once a sufficiently hot core wasestablished. Two of the three tested hearths (Houses 3 and 8)contain chemical traces of both disordered calcite and carbonatehydroxylapatite, as well as fragments of wood charcoal and burnedbone (Fig. 3). Calcite in the hearths, based on the n2/n4 ratios, ispyrogenic. Carbonate hydroxylapatite in hearth soils was derivedfrom burned bone fragments and powder, not from the dissolutionand reprecipitation of ash calcite (Schiegl et al., 2003). Ash calcite isstable in the dwelling deposits, indicating it has not undergonetransformation into carbonate hydroxylapatite. The carbonatehydroxylapatite identified in the hearth samples was derived fromthe calcined bone powder produced when bone is burned at hightemperatures (Stiner et al., 1995). Calcined bone fragments werefurther powdered and incorporated into the soil matrix via cry-oturbation. The absence of material and chemical evidence fora hearth in House 2 likely represents cleaning activity, which, alongwith the charcoal identified in several cores, supports that burnedmatter was dumped across the site.
Fuel selection is commonly guided by the availability of mate-rials. A viable supply of wood is lacking on the Sub-Arctic tundra,but, during large scale fall caribou hunting, caribou bone providedan alternative fuel source. Crystallinity index, C/P values, andorganic matter, collagen, and phosphate peak changes for exca-vated caribou bone indicate high intensity burning (Fig. 4; Table 2).Unburned samples from dwelling floors were well preserved,demonstrating the diagenetic alterations recorded for hearthsamples were caused primarily by burning. The appearance ofsharp peaks at 632, 1090, and 962 cm�1 in the spectra with thehighest CI values demonstrate that some samples are calcined andwere burned at temperatures exceeding 700 �C (Thompson et al.,2009; Weiner, 2010) (Fig. 4). An experimental heating study ina stállo pit-house reconstruction demonstrated that in outdoorconditions around�10 �C, a birchwood firewith a core temperatureof 437.7 �C produces an ambient indoor temperature rangingbetween 13.2 and 26.9 �C (Liedgren and Östlund, 2011). Hearthsused in the tested Ikirahak dwellings reached higher core temper-atures, indicating the houses were well heated, which would not benecessary duringwarmermonths. High temperatureswere requiredbecause the dwellings were occupied during cold seasons. Main-taining hearths at temperatures exceeding 700 �C would have beenan inefficient use of limited fuel during warmer months.
There is no evidence for the use of outdoor hearths. The spatialextent of montgomeryite suggests authigenic formation fromwidespread refuse, not from ash in discrete, primary hearth areas.There are traces of charcoal in some soils from across the site, butoverall, the soils are lacking disorganized calcite, carbonatehydroxylapatite, and hematite. Moreover, Hodgetts et al.’s (2011)magnetometer survey of the site did not detect any signs ofburning outside of dwelling contexts. It would have been an inef-ficient use of sparse fuel to maintain outdoor fires during colderseasons, as these hearths would have been poorly protected againstwind and moisture, and they would have required significantlymore fuel to produce a sustainable amount of heat. Hearth featureswere only established inside dwellings, supporting a cold seasonoccupation and providing insight into the use of household ratherthan communal economic units. Using the frequencies and sizes ofhearths and middens to define socio-economics, however, firstrequires an understanding of feature contemporaneity.
Bone burning, widespread chemical alterations of soils, invest-ments into building pit-houses, and excavated hunting and hide
processing equipment at Ikirahak indicate low residential mobilityand repeated occupation. We propose that this tundra-based sitefunctioned as a late fall caribou hunting base camp and processingfacility, which is a type of settlement behaviour considerablydifferent from that documented within the Beverly Caribou Range.It also diverges radically from patterns documented for their Chi-pewyan descendents, who visited the tundra in small, highlymobile groups during summer (Hearne, 1958). Our model contendsthat southbound caribou herds would have been intercepted atwater crossings, such as those identified around Ikirahak Island,during early and mid fall to build surpluses of hides and meat.Housing and clothing a group of families would take severalhundred hides, and given the thin, supple nature of caribou leather,people would have to replace their products regularly (Irimoto,1981; Thompson, 1994). Hides were in the best condition forthese purposes during fall, after warble fly damage healed, whenaggregated caribou were moving south (Hearne, 1958). Sites likeIkirahakwere not rapidly abandoned immediately after the huntingseason. People stayed into the late fall/early winter to dry meat andprepare hides. Fresh meat would have supported the group duringtheir stay at the settlement, while dried meat would have beenprepared and stored for the long trip to the southern forest and forsupplementing diets throughout winter. We are currently investi-gating the impacts of meat and hide processing and storage usingx-ray fluorescence, inductively coupled plasma mass-spectroscopy,gas chromatography, and FTIR analyses of soils and referencematerials.
6. Conclusions
This research contributes to demonstrating that FTIR is anaccessible, rapid, and cost-effective part of themicro-archaeologicaltool-kit, and that it is valuable in reconstructions of functions andorganizations of spaces at hunter-gatherer settlements. The rapidityand low cost of analyses make the technique useful for identifyingareas on sites for further, more complicated and expensive analyses.In our case study, there are no spatially patterned concentrations ofauthigenic phosphates, supporting the argument that faunalelements from spatially organized middens have not been buried,completely decayed, and reprecipitated. Excavated unburned bonesarewell preserved, verifying that a buried bonemiddenwould likelysurvive in this environment. We argue that disordered mont-gomeryite formed from the diagenesis and reprecipitation ofcarbonates, phosphates, andmagnesiumadded to the systemby thedispersal of ash and caribou processing residues. Additionalresearch using experimental deposits, lysimetry, micromorphology,and electron microprobe will help clarify the formation processesand behaviour of authigenic phosphates on open-air hunter-gatherer sites.
Our results support the use of CI, C/P, calcite, and carbonatehydroxylapatite for distinguishing fuel sources in hunter-gathererhearths. Several excavated bones exhibit high intensity burning,supporting the arguments that bone was used as fuel and that thesite was occupied during a cold season. Additional experimentalresearch concerning bone burning/decomposition and mineraliza-tion specifically in caribou bone is necessary, given that ratios oforganic and mineral components vary across species, age, andhabitat. Such variations may influence measurements of crystal-linity and burning. Infrared spectra also have signatures for disor-ganized calcite and carbonate hydroxylapatite in two of the threetested pit-house hearths, indicating the use of both wood and bonefor fuel. Based on the absence of hematite, disordered calcite, andcarbonate hydroxylapatite in soils from across the settlement,hearth features were exclusive to dwellings. The use of semi-subterranean dwellings at the Ikirahak site, a housing type
previously unidentified in Taltheilei archaeological record, suggeststhat people occupied this area during a cold season. Visible andchemical traces of burned bone in hearth features support thisargument. Bone provided a viable fuel alternative on the tree-lesstundra during late fall stays, which we argue is one of the primaryreasons the Ikirahak site lacks a bone midden. A late fall occupationof a tundra-based site indicates that Taltheilei settlement strategieswere more dynamic than previous research has recognized.
Acknowledgements
We thank the Social Sciences and Humanities Research Councilof Canada, the Department of Indian and Northern Affairs Canada,International Polar Year Canada, and the University of CalgaryDepartment of Archaeology for funding our project. We thankWade White from the University of Calgary Department of Chem-istry for FTIR training, Derek Wilson from the University of CalgaryDepartment of Geography for PSA training, Calla McNamee andHoward Cyr from the University of Calgary Department ofArchaeology for their micromorphology research, and Sean Pick-ering from the University of Calgary Department of Archaeology forleading excavations and lithic analyses. We also thank ourreviewers for their helpful comments.
References
Albert, R.M., Bar-Yosef, O., Meignen, L., Weiner, S., 2003. Quantitative phytolithstudy of hearths from the Natufian and Middle Palaeolithic levels of HayonimCave (Galilee, Israel). Journal of Archaeological Science 30, 461e480.
Aspinall, A., Gaffney, C., Schmidt, A., 2008. Magnetometry for Archaeologists.Altamira Press, Lanham.
Bamforth, D.B., Becker, M., Hudson, J., 2005. Intrasite spatial analysis, ethno-archaeology, and Paleoindian land-use on the Great Plains: the Allen Site.American Antiquity 70, 561e580.
Beck, M.E., Hill Jr., M.E., 2004. Rubbish, relatives, and residence: the family use ofmiddens. Journal of Archaeological Method and Theory 11, 297e333.
Berna, F.,Matthews,A.,Weiner, S., 2004. Solubilitiesofbonemineral fromarchaeologicalsites: the recrystallization window. Journal of Archaeological Science 31, 867e882.
Berna, F., Behar, A., Shahack-Gross, R., Berg, J., Boaretto, E., Gilboa, A., Sharon, I.,Shale, S., Shilstein, S., Yahalom-Mack, N., Zorn, J.R., Weiner, S., 2007. Sedimentsexposed to high temperatures: reconstructing pyrotechnological processes inLate Bronze and Iron Age Strata at Tel Dor (Israel). Journal of ArchaeologicalScience 34, 358e373.
Bhargava, R., Levin, I.W., 2005. Spectrochemical Analysis Using Infrared Multi-channel Detectors. Blackwell, Oxford.
Binford, L.R., 1983. In Pursuit of the Past: Decoding the Archaeological Record.Thames and Hudson, New York.
Binford, L.R., 1990. Mobility, housing and environment: a comparative study. Jour-nal of Anthropological Research 46, 119e152.
Butler, D.H., 2011. Exploring soilscapes and places inside labrador Inuit winterdwellings. Canadian Journal of Archaeology 35, 55e85.
Chu, V., Regev, L., Weiner, S., Boaretto, E., 2008. Differentiating between anthro-pogenic calcite in plaster, ash and natural calcite using infrared spectroscopy:implications in archaeology. Journal of Archaeological Science 35, 905e911.
Cronyn, J.M., 1990. The Elements of Archaeological Conservation. Routledge,London.
Dawson, P., Pickering, S., McNamee, C., Cyr, H., Hodgetts, L., 2009. Permit Report2008_023A: a Report on Archaeological Fieldwork at Ikirahak, Maguse Lake,Nunavut. Report on File at the Department of Culture, Language, Elders, andYouth, Iqaluit.
Goldberg, P., Nathan, Y., 1975. The phosphate mineralogy of et-Tabun cave, MountCarmel, Israel. Mineralogical Magazine 40, 253e258.
Goldberg, P., Dibble, H., Berna, F., Sandgathe, D., McPherron, S.J.P., Turq, A., 2012.New evidence on Neandertal use of fire: examples from Roc de Marsal and Pechde l’Azé IV. Quaternary International 247, 225e240.
Gordon, B.C., 1975. Of Men and Herds in Barrenland Prehistory. In: Mercury Series,Archaeological Survey of Canada, 28. National Museums of Canada, Ottawa.
Gordon, B.C., 1996. People of Sunlight, People of Starlight. In: Mercury Series,Archaeological Survey of Canada, 154. Canadian Museum of Civilization, Hull.
Haynes, S., Searle, J., Bretman, A., Dobney, K., 2002. Bone preservation and ancientDNA: the application of screening methods for predicating DNA survival.Journal of Archaeological Science 29, 585e592.
Hearne, S., 1958. A Journey from Prince of Wales Fort in Hudson’s Bay to theNorthern Ocean: 1769, 1770, 1771, 1772. The Macmillan Company of Canada,Toronto.
Hedges, R.E.M., 2002. Bone diagenesis: an overview of processes. Archaeometry 44,319e328.
Helwig, K., Monahan, V., Poulin, J., 2008. The identification of hafting adhesive ona slotted antler point from a Southwest Yukon ice patch. American Antiquity 73,279e288.
Hodgetts, L., Dawson, P., Eastaugh, E., 2011. Archaeological magnetometry in anArctic setting: a case study from Maguse Lake, Nunavut. Journal of Archaeo-logical Science 38, 1754e1762.
Huaqing, L., 1989. Effect of chemical adsorption of calcium oxalate monohydrate onDTA baseline. Thermochimica Acta 141, 151e157.
Irimoto, T., 1981. Chipewyan Ecology: Group Structure and Caribou Hunting System.In: Senri Ethnological Studies No. 8. National Museum of Ethnology, Osaka.
Karkanas, P., Kyparissi-Apostolika, N., Bar-Yosef, O., Weiner, S., 1999. Mineralassemblages in Theopetra, Greece: a framework for understanding diagenesisin a prehistoric cave. Journal of Archaeological Science 26, 1171e1180.
Karkanas, P., Bar-Yosef, O., Goldberg, P., Weiner, S., 2000. Diagenesis in prehistoriccaves: the use of minerals that form in situ to assess the completeness of thearchaeological record. Journal of Archaeological Science 27, 915e929.
Karkanas, P., Rigaud, J.-P., Simek, J., Albert, R., Weiner, S., 2002. Ash, bones andguano: a study of the minerals and phytoliths in the sediments of Grotte XVI,Dordogne, France. Journal of Archaeological Science 29, 721e732.
Karkanas, P., Shahack-Gross, R., Ayalon, A., Bar-Matthews, M., Barkai, R., Frumkin, A.,Gopher, A., Stiner, M.C., 2007. Evidence for habitual use of fire at the end of theLower Paleolithic: site-formation processes at Qesem Cave, Israel. Journal ofHuman Evolution 53, 197e212.
Kelly, R.L., Poyer, L., Tucker, B., 2005. An ethnoarchaeological study of mobility,architectural investment, and food sharing among Madagascar’s Mikea. Amer-ican Anthropologist 107, 403e416.
Kent, S., 1999. The archaeological visibility of storage: delineating storage from trashareas. American Antiquity 64, 79e94.
Kimmel Center for Archaeological Science (KCAS) Infrared Spectra Catalogue. http://www.weizmann.ac.il/kimmel-arch/islib.html.
Knudson, K.J., Frink, L., 2010. Ethnoarchaeological analysis of Arctic fish processing:chemical characterization of soils on Nelson Island, Alaska. Journal of Archae-ological Science 37, 769e783.
Lebon, M., Reiche, I., Bahain, J.J., Chadefaux, C., Moigne, A.M., Fröhlich, F., Sémah, F.,Schwarcz, H.P., Falguères, C., 2010. New parameters for the characterization ofdiagenetic alterations and heat-induced changes of fossil bone mineral usingFourier transform infrared spectrometry. Journal of Archaeological Science 37,2265e2276.
LeMouel, J.F., LeMouel, M., 2001. Aspects of early Thule culture as seen in thearchitecture of a site on Victoria Island, Amundsen Gulf area. Arctic 55, 167e189.
Liedgren, L.G., Östlund, L., 2011. Heat, smoke and fuel consumption in a highmountain stállo-hut, northern Sweden: experimental burning of fresh birchwood during winter. Journal of Archaeological Science 38, 903e912.
McGuire, R.H., Schiffer, M.B., 1983. A theory of architectural design. Journal ofAnthropological Archaeology 2, 227e303.
McNamee, C., Cyr, H., and Dawson, P., 2009. Site Formation Processes of Sub ArcticPithouses, Ikirahak Island, Nunavut. Poster Presented at the 2009 Society forAmerican Archaeology Conference.
Middleton, W.D., Price, T.D., 1996. Identification of activity areas by multi-elementcharacterization of sediments from modern and archaeological house floorsusing inductively coupled plasma-atomic emission spectroscopy. Journal ofArchaeological Science 23, 673e687.
Munro, L.E., Longstaffe, F.J., White, C.D., 2007. Burning and boiling of modern deerbone: effects on crystallinity and oxygen isotope composition of bioapatitephosphate. Palaeogeography, Palaeoclimatology, Palaeoecology 249, 90e102.
Nagy, G., Lorand, T., Patonai, Z., Montsoko, G., Bajnoczky, I., Marcsik, A., Mark, L.,2008. Analysis of pathological and non-pathological human skeletal remains byFT-IR spectroscopy. Forensic Science International 175, 55e60.
Nash, R.J., 1970. The Prehistory of northern Manitoba. In: Hlady, W.M. (Ed.), TenThousand Years: Archaeology in Manitoba. Manitoba Archaeological Society,Altona, pp. 76e92.
Nielsen-Marsh, C.M., Hedges, R.E.M., 1999. Bone porosity and the use of mercuryintrusion porosimetry in bone diagenesis studies. Archaeometry 41, 165e174.
Nriagu, O.J., 1976. Phosphate-clay mineral relations in soils and sediments. Cana-dian Journal of Earth Sciences 13, 717e736.
Oetelaar, G.A., 1993. Identifying site structure in the archaeological record: an Illi-nois Mississippian example. American Antiquity 58, 662e687.
Oswald, D., 1984. The use of site structure to infer social and technological orga-nization from the archaeological record. In: Hall, M. (Ed.) Frontiers: Southern
African Archaeology Today. British Archaeological Reports 207. British Archae-ological Reports, Oxford, pp. 297e309.
Petch, V. 1992. An Archaeological Survey of the Thlewiaza River from Fort Hall Laketo Nuellin Lake. Report on File with Manitoba Natural Resources, Parks Branch.
Petch, V., Punter, E., Groom, H., Bahr, J., Wilson, R., 1997. The Robertson Esker inNorthern Manitoba: a Journey Through Time. Northern Lights HeritageServices, Winnipeg.
Regev, L., Poduska, K.M., Addadi, L., Weiner, S., Boaretto, E., 2010. Distinguishingbetween calcites formed by different mechanisms using infrared spectrometry:archaeological applications. Journal of Archaeological Science 37, 3022e3029.
Rendon, J.L., Serna, C.J., 1981. IR spectra of powder hematite: effects of particle sizeand shape. Clay Minerals 16, 375e381.
Shahack-Gross, R., Albert, R.M., Gilboa, A., Nagar-Hilman, O., Sharon, I., Weiner, S.,2005. Geoarchaeology in an urban context: the uses of space in a Phoenicianmonumental building at Tel Dor (Israel). Journal of Archaeological Science 32,1417e1431.
Shipman, P., Foster, G., Schoeninger, M., 1984. Burnt bones and teeth: an experi-mental study of color, morphology, crystal structure and shrinkage. Journal ofArchaeological Science 11, 307e325.
Squires, K.E., Thompson, T.J.U., Islam, M., Chamberlain, A., 2011. The application ofhistomorphometry and Fourier Transform Infrared Spectroscopy to the analysisof early Anglo-Saxon burned bone. Journal of Archaeological Science 38, 2399e2409.
Stiner, M., Kuhn, S., Weiner, S., Bar-Yosef, O., 1995. Differential burning, recrystal-lization, and fragmentation of archaeological bone. Journal of ArchaeologicalScience 22, 223e237.
Stiner, M., Kuhn, S., Surovell, T., Goldberg, P., Meignen, L., Weiner, S., Bar-Yosef, O.,2001. Bone preservation in Hayonim Cave (Israel): a macroscopic and miner-alogical study. Journal of Archaeological Science 28, 643e659.
Surovell, T.A., Stiner, M.C., 2001. Standardizing infra-red measures of bone mineralcrystallinity: an experimental approach. Journal of Archaeological Science 28,633e642.
Tatzber, M., Stemmer, M., Spiegel, H., Katzlberger, C., Haberhauer, G.,Gerzabek, M.H., 2007. An alternative method to measure carbonate in soils byFT-IR spectroscopy. Environmental Chemical Letters 5, 9e12.
Thompson, J., 1994. From the Land: Two Hundred Years of Dene Clothing. CanadianMuseum of Civilization.
Thompson, T.J.U., Gauthier, M., Islam, M., 2009. The application of a new method ofFourier Transform Infrared Spectroscopy to the analysis of burned bone. Journalof Archaeological Science 36, 910e914.
Thompson, T.J.U., Islam, M., Piduru, K., Marcel, A., 2010. An investigation into theinternal and external variables acting on crystallinity index using FourierTransform Infrared Spectroscopy on unaltered and burned bone. Palae-ogeography, Palaeoclimatology, Palaeoecology 299, 168e174.
Trueman, C.N., Privat, K., Field, J., 2008. Why do crystallinity values fail to predictthe extent of diagenetic alteration of bone mineral? Palaeogeography, Palae-oclimatology, Palaeoecology 266, 160e167.
Weiner, S., 2010. Microarchaeology. Beyond the Visual Archaeological Record.Cambridge University Press.
Weiner, S., Price, P.A., 1986. Disaggregation of bone into crystals. Calcified TissueInternational 39, 365e375.
Weiner, S., Bar-Yosef, O., 1990. States of preservation of bones from prehistoric sitesin the Near East: a survey. Journal of Archaeological Science 17, 187e196.
