lable at ScienceDirect

Journal of Archaeological Science 40 (2013) 4576e4585

Contents lists avai

Journal of Archaeological Science

journal homepage: http: / /www.elsevier .com/locate/ jas

Paleodiet in northern Chile through the Holocene: extremely heavyd15N values in dental calculus suggest a guano-derived signature?

Simon R. Poulson a,*, Susan C. Kuzminsky b, G. Richard Scott c, Vivien G. Standen d,Bernardo Arriaza d,e, Iván Muñoz d, Lindsay Dorio c

aDepartment of Geological Sciences and Engineering/MS-172, University of Nevada Reno, 1664 N. Virginia St., Reno, NV 89557, USAbDepartment of Anthropology, University of California, Santa Barbara, CA 93106-3210, USAcDepartment of Anthropology/MS-096, University of Nevada Reno, Reno, NV 89557, USAdDepartamento de Antropología, Universidad de Tarapacá, Arica, Chilee Instituto de Alta Investigación, Universidad de Tarapacá, Arica, Chile

a r t i c l e i n f o

Article history:Received 26 October 2012Received in revised form5 July 2013Accepted 7 July 2013

Keywords:BioarchaeologyPaleodietStable isotopesDental calculusChile prehistoryNitrogenCarbonSeabird guano

* Corresponding author. Tel.: þ1 775 784 1104; faxE-mail address: poulson@mines.unr.edu (S.R. Poul

0305-4403/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.jas.2013.07.009

a b s t r a c t

Dental calculus extracted from the teeth of 28 prehistoric human skeletons from coastal and valleyarchaeological sites in northern Chile that date from the Archaic period (w2300 BC) to the Late Inter-mediate period (AD 1476) was analyzed for stable isotope (d15N, d13C) compositions. d15N compositionswere extremely heavy (þ17.8 to þ33.1&), but comparable to some studies using conventional humanbiomaterials (hair, nail, muscle) at other prehistoric sites in northern Chile. There was a negative cor-relation between d15N and d13C for coastal sites, but a positive correlation for valley sites. Results for thevalley sites point to a diet that was influenced by marine resources throughout all time periods. Theunusually heavy d15N values for the coastal sites require a dietary component with a d15N compositionsignificantly heavier than that of marine resources. The hyper-aridity at the study area (mean annualrainfall of 0.5e0.6 mm/year) is a likely contributing factor to the production of heavy d15N values, but isunlikely to account for the heaviest d15N values (>þ30&). One possible explanation for the heaviest d15Nvalues is that dietary components were impacted by isotopically-heavy guano, which is abundant in theregion. Guano may have been used as a fertilizer during crop cultivation at the onset of the Formativeperiod, continuing through the Historic period after Spanish contact. The indirect impact of guano fromabundant bird habitats in the region may have influenced wild foodstuffs harvested throughout the pre-agricultural period. Results provide support for the utility of dental calculus as a proxy for obtainingstable carbon and nitrogen isotope signatures for use in paleodietary studies.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Stable carbon and nitrogen isotopes (d15N and d13C) have longbeen used to qualify and/or quantify dietary compositions forhumans and other organisms because they are an effective deter-minant of possible food sources (e.g. Schoeninger, 2009, 2010;Schwarcz and Schoeninger, 1991). Stable isotope analysis of allsample material types is destructive, consumes a finite amount ofsample material, and may not be allowed when laws designed toprotect cultural remains, such as the Native American Graves Pro-tection and Repatriation Act in the United States, preclude thedestruction of primary biomaterials (i.e. bone, hair, nails, muscle

: þ1 775 784 1833.son).

All rights reserved.

tissue). In a recent study, Scott and Poulson (2012) suggested thatdental calculus might serve as a new proxy for estimating stablecarbon and nitrogen isotope ratios with application to paleodietarystudies. Although stable isotope analysis of dental calculus is alsodestructive, consideration of calculus as a secondary (or “add-on”)biomaterial rather than a primary biomaterial may alleviate cura-torial concerns regarding the destruction of sample material insome cases.

Dental calculus is a biomineral derived from plaque, a biofilmthat forms on tooth surfaces as a byproduct of microbial activity inthe mouth (Scott and Poulson, 2012). If this plaque biofilm is notremoved, it hardens to form calculus after approximately ten days.Calculus continues to accrete onto teeth over the course of an in-dividual’s lifetime. In the absence of modern dental hygiene prac-tices that remove plaque and calculus, the presence of dentalcalculus in prehistoric teeth is a common occurrence. In contrast to

Fig. 1. Location of samples from sites in Arica and the Azapa Valley, Chile.

S.R. Poulson et al. / Journal of Archaeological Science 40 (2013) 4576e4585 4577

other body tissues (e.g. collagen, muscle tissue, hair, fingernails)that have been used for stable isotope analysis, and consist ofproteins with C/N weight ratios that lie within a relatively narrowrange (e.g. 2.9e3.6 for collagen, 2.9e3.8 for hair; DeNiro, 1985;O’Connell and Hedges, 1999), calculus is not a conventional ‘tis-sue’ with a fixed C/N weight ratio. Mineralogical investigation re-veals that calculus is composed primarily of various calciumphosphate minerals, including brushite, whitlockite, octacalciumphosphate, and hydroxyapatite (Hayashizaki et al., 2008; White,1997), although organic components such as proteins, carbohy-drates, lipids, and glycoproteins comprise 15e20% of the dryweightof calculus (Lieverse, 1999), and these organic components provideenough carbon and nitrogen to facilitate stable isotope analysis(d13C and d15N). To the best of our knowledge, only one previousstudy (Scott and Poulson, 2012) has analyzed the stable isotopiccomposition of dental calculus.

Stable isotope analysis of dental calculus was performed for asuite of samples from the desert region of northern Chile toinvestigate paleodiet. The study area was selected for its abundantskeletal collections, excellent preservation, and broad temporalsequence, which covers the Archaic period (w2300 BC) to the endof the Late Intermediate period (AD 1476), just prior to theexpansion of the Inca Empire into the region. In contrast to theEuropean study by Scott and Poulson (2012), the later prehistoricpopulations in this desert region of northern Chile subsisted onmaize, a C4 plant, and utilized marine resources intensively (MuñozOvalle and Focacci, 1985; Santoro et al., 2003; Watson et al., 2011).

2. Environment e the extreme north of Chile

The extreme north of Chile is broadly divided into threeecological zones: (1) the Pacific coastal zone; (2) the intermediatezone with oases and valleys that transect the arid Atacama Desert;and (3) the puna, precordillera, and altiplano, which make up theAndean highlands (Rivera, 1991). Microclimates found within thesezones foster some plant and animal diversity, but human survival innorthern Chile during prehistoric times was difficult due to thehyper-arid environment (see Bryson et al., 2001 for paleoclimaticevidence). During the prehistoric period, human survival dependedupon settling in valleys and valley openings where fresh watersources were located. Despite the dearth of terrestrial plants andanimals in the Atacama Desert, the cold waters of the PacificHumboldt Current supported an abundance of marine resources(Arriaza et al., 2008; Llagostera, 1979). These microclimates andecological zones were interconnected throughout prehistory,providing groups with extensive trade networks and access tonatural resources (Rivera, 1991).

The study area is the modern city of Arica and the adjacentAzapa Valley, just a few kilometers south of the modern politicalborder with Peru (Fig. 1). The coastal city of Arica and the AzapaValley have extensive, well-preserved archaeological deposits thatappear as early as 7000 BC (Standen et al., 2004). Scholars haveargued that mild annual temperatures, fresh water supply inneighboring valleys, and diverse marine and bird life made this anattractive area for human occupation in an otherwise hostileenvironment (Arriaza, 1995a, 1995b; Arriaza et al., 2008; Brysonet al., 2001; Dillehay, 2008).

Throughout the Archaic period, people who lived on theextreme north coast of Chile were thought to have relied heavily onmarine resources (Arriaza et al., 2008, 2010; Aufderheide et al.,1993; Standen et al., 2004). Recent research indicates that theseearly coastal populations were not fully nomadic, but instead, livedin year-round camps along the coast (Arriaza et al., 2008).Following the Archaic period, settlements increased in size at thecoast, while people began to form permanent settlements in the

Azapa Valley (Muñoz Ovalle, 2004; Sutter and Mertz, 2004). Plantdomestication began in the region at approximately 1500 BC (thebeginning of the Formative period), possibly as a response triggeredby a climatic shift (Bryson et al., 2001; Marquet et al., 2012). Duringthe Middle Horizon (AD 750e1100), the Azapa Valley and adjacentcoast experienced further regional development, influenced by theTiwanaku culture, increased village formation, and more extensiveirrigation systems (Rivera, 1991). The Middle Horizon also appearsto have been a time of increased textile and pottery production andtrade route expansion between the coast, valleys, and highlandzones (Rivera, 1991; Roberts et al., 2013). The Late Intermediateperiod (AD 1100e1476) is characterized by continued exchangenetworks between ecological zones and the emergence of regionalchiefdoms with ceramic styles from the San Miguel (1100e1200AD) and Gentilar (wAD 1400) cultures who inhabited coast andvalley sites (Sutter and Mertz, 2004). The final Prehispanic culturalsequence at coast and valley sites was dominated by the expansionof the Inca Empire (wAD 1500) into Chile, which led to thedevelopment of outposts and roads that facilitated access to min-eral resources, continued craft production, and exchange of goodsthroughout the region (Rivera, 2008; Roberts et al., 2013).

Several dietary patterns are apparent at the coastal and AzapaValley sites throughout the prehistoric period. Evidence of a heavilymarine-based diet through the Archaic period is supported by gooddental health and intestinal parasites associated with marine foodconsumption (Arriaza,1995a; Arriaza et al., 2008; Kelley et al., 1991;Reinhard and Urban, 2003). After the Archaic period, people tran-sitioned from their near-exclusive dependence on ocean foods to amixed diet of cultivates, marine foods, and domesticated animals(Bryson et al., 2001; Rivera, 1991; Watson et al., 2010). Bio-archaeological research conducted byWatson et al. (2010) revealedmore dental decay among Azapa Valley inhabitants during theFormative period, consistent with an increase in cariogenic foods,such as maize. However, despite the introduction of tubers,

S.R. Poulson et al. / Journal of Archaeological Science 40 (2013) 4576e45854578

peppers, maize, beans, and squash into the Azapa Valley, severalstudies indicate that marine resources continued to be an impor-tant dietary staple throughout the prehistoric period, and peoplemay have increased ocean harvests, rather than decreased them(Roberts et al., 2013; Santoro et al., 2003; Standen et al., 1997;Watson et al., 2011). Archaeological data from the Azapa Valleysupport this scenario, where agropastoral evidence was found withfishhooks, fishing nets, and other items associated with theacquisition of marine foods (Muñoz Ovalle and Focacci, 1985). Onereason for the continued reliance on marine foods may have beenthat they were less severely affected by fluctuating climate change(Rothhammer and Santoro, 2001). With the Azapa Valley locatedonly 18 km from the coast, it would have been relatively easy forpeople to take advantage of abundant marine protein sources. Asvillage life, trade networks, and political integration intensifiedthrough the Middle Horizon and later periods, larger populationsled to the demand for food surpluses and increased agro-pastoralism (Rivera, 1991; Roberts et al., 2013). On the other hand,archaeoparasitological research conducted by Santoro et al. (2003)in the nearby Lluta Valley showed that the Inca expansion duringthe Late period led not only to political changes in the region, but anabnormally high prevalence of human parasites, including thoseassociated with fish consumption.

Given the archaeological and biological data reported for thestudy area, the earliest samples were expected to show isotopevalues consistent with a diet rich in marine foods. Individuals fromcoastal sites, with easier access to marine foods, should showheavier d15N values than valley sites through all time periods. Asthe introduction of agriculture and animal domestication spreadthroughout the coast, valleys, and highlands during the Formativeperiod, the isotope results should reflect the increased influence ofcarbohydrates (e.g. maize and tubers) and terrestrial meats (e.g.camelid). Bioarchaeological data (e.g. Aufderheide et al., 1993,1994;Roberts et al., 2013; Standen et al., 1997; Watson et al., 2011) havedemonstrated a continued reliance on marine foods through alltime periods. As such, isotope results were expected to indicate thata mixed subsistence economy was utilized at valley sites, particu-larly during the Middle Horizon and Late Intermediate periodswhen agropastoralism and trade networks intensified to supportlarger populations (Rivera, 1991, 2008).

3. Materials and methods

Thirty-five calculus samples were obtained from 28 skeletonscurated at the Museo Arqueológico Universidad de Tarapaca thatare associated with seven coastal and valley archaeological sites inArica and the Azapa Valley (Table 1, Fig. 1). Individuals from thethree coastal sites (Morro 1-6, Quiani 7, and Playa Miller 7) datefrom the Archaic period (2300e1800 BC) to the Formative period(500 BCeAD 500). Individuals from the four Azapa Valley sites (AZ-6, AZ-8, AZ-75, and AZ-140) date from the Formative period (500BCeAD 500) to the Late Intermediate period (ca. AD 1000e1476).The samples are associated with prehistoric people whose subsis-tence economies centered on hunting and gathering, fishing,shellfish-collecting, pastoralism, and agriculture. Such practiceswere varied and changed over time in the coastal and valley areas.

Individuals with moderate to heavy calculus buildup on theincisors or molars of the maxilla and mandible were selected forsampling. Following the protocol outlined in Scott and Poulson(2012), approximately 5e10 mg of calculus was scraped off thebuccal and/or lingual sides of a tooth. In some cases the calculuswas removed without the aid of a tool, while stubborn calculusdeposits were carefully scraped away with a dental pick.

As a second level of analysis, calculus was removed from twodifferent areas of the mandible and maxilla for seven of the 28

individuals to determinewhether isotope values would be similar ifsamples were tested from different areas of the mouth. For theseindividuals, calculus was removed from the buccal/lingual sides ofthe lower incisors and the buccal/lingual sides of upper or lowercheek teeth. In some cases, calculus had to be removed from morethan one tooth to collect sufficient sample material.

Calculus was rinsed with distilled, de-ionized water, dried at50 �C, and then powdered in a steel mortar and pestle. Approxi-mately 1.5e8 mg of calculus was used for each analysis. Stableisotope analyses and elemental concentrations were performedusing a Eurovector elemental analyzer interfaced to an Isoprimestable isotope ratio mass spectrometer, after the method ofWerner et al. (1999), which provides for bulk analyses (i.e. analysesof all C and N present) of weight % C, weight % N, d13C, and d15N.Stable isotope results are reported in units of & in the usuald notation vs. VPDB for carbon, and vs. air for nitrogen. Duplicateor triplicate analyses (i.e. multiple analysis of material from asingle powdered sample) were performed for sixteen samples andresulted in a mean reproducibility (one standard deviation, whichincludes any sample heterogeneity that may be present) of 0.1&for d13C, 0.1& for d15N, 0.15% for weight % C, and 0.02% for weight %N. Acetanilide (Costech Analytical Technologies Inc., Valencia, CA)was used as a standard for isotope and elemental concentrationanalyses.

Statistical analyses were conducted with SPSS version 21.0 tocompare the mean isotope values between the site locations (coastvs. valley) and time periods (Archaic to Formative vs. Middle Ho-rizon to Late Intermediate periods). To improve the uneven distri-bution of the sample for each of the four time periods (Table 1), theArchaic and Formative periods and the Middle Horizon and LateIntermediate periods were combined to form early and latecomparative groups. A ShapiroWilk test of normality indicated thatthe data used in the analyses followed a normal distribution for thecoast vs. valley and for the Archaic/Formative vs. the Middle Hori-zon/Late Intermediate period comparisons. These comparative datawere then subjected to a Levene’s test of the equality of variances,followed by an independent samples t-test using the Welch’scorrection that assumes unequal variances. Effect sizes are reportedwith Cohen’s d and Pearson’s r.

4. Results

Stable isotope (d13C and d15N) and elemental concentration(weight % C, weight % N, C/N weight ratio) results for all dentalcalculus samples are presented in Table 1, including mean valuesand values of 1 standard deviation (1s) for each site, for coastal vs.valley sites, and for all samples. For all samples, weight % C rangedfrom 5.97 to 37.20%, with a mean of 11.74% and a standard deviationof 5.50%; weight % N ranged from 1.05 to 2.42%, with a mean of1.41% and a standard deviation of 0.26%. There were no systematicdifferences in elemental concentrations between coast and valleysites or any systematic relationship between elemental concen-trations and early (Archaic to Formative) vs. late (Middle Horizon toLate Intermediate) periods.

For stable isotope compositions, d15N ranged from þ17.8to þ33.1&, with a mean of þ24.1& and a standard deviation of4.3&. Calculus d13C vs. d15N is plotted in Fig. 2, with coast and valleysamples plotted separately. An independent samples t-test indi-cated that coastal d13C values (�18.8 � 1.0&) were marginallylower than valley d13C values (�17.4 � 1.7&: t ¼ 2.045, df ¼ 16,p ¼ .06, d ¼ 0.80). However, a t-test (t ¼ 6.478, df ¼ 8, p < .001,d ¼ 2.92) revealed a large difference between mean coastal d15Nvalues (þ30.4� 2.3&) andmean valley d15N values (þ22.3� 2.6&),a difference of almost three standard deviations. There was nosystematic relationship between either d13C or d15N vs. early or late

Table 1Site name, cultural group, location, period, specimen numbers, teeth sampled, d15N, wt% N, d13C, and wt% C for all samples. Paired and shaded values represent separatesamples from the same individual. 1Focacci and Chacon (1989); 2Dauelsberg (1974); 3Focacci (1974); 4Aufderheide et al. (2002); 5Focacci (1990); 6Sutter and Mertz (2004).

S.R. Poulson et al. / Journal of Archaeological Science 40 (2013) 4576e4585 4579

0

10

20

30

40

50

-30 -25 -20 -15 -10 -5

δ15

N (‰

)

δ13

C (‰)

seabird guano-δ15N

Fig. 2. Arica e coast, and Azapa e valley: human dental calculus analyses from northern Chile (this study). Pisagua, Chile e human hair and muscle tissue from mummies:Aufderheide et al. (1994). Osmore Valley, Peru e coastal sites: human bone collagen from Roca Verde, Wawakiki, and San Geronimo sites (Tomczak, 2001, 2003). Caleta Vitor ecoast: human bone collagen from northern Chile (Roberts et al., 2013). Southern California, USA e human bone collagen, various sites in the Santa Barbara Channel area (Walker andDeNiro, 1986). Neo and paleo-Aleuts, eastern Aleutians, Alaska, USA e average values for human bone collagen (Byers et al., 2011). South Pacific fish e modern South Pacific fish(Tieszen and Chapman, 1992). Pisagua, Chile e plant grave goods: seaweed, maize cobs, distichlis, and plant stems (Aufderheide et al., 1994). Prehistoric plants, Peru - plants frommuseum collections, sampled from various sites in Peru (DeNiro and Hastorf, 1985: data for C3 plants, C4 plants and legumes from Table 2 of DeNiro and Hastorf, 1985; that had notreceived any thermal, chemical, or microbial treatments). Guano-fertilized maize - growth chamber experiments for G1 and G2 guano fertilizer treatments at 115 days, guano-d15N ¼ þ26.7& (Szpak et al., 2012a). Seabird guano-d15N e range of d15N values (only) for guano, compiled in Table 1 of Szpak et al. (2012a).

S.R. Poulson et al. / Journal of Archaeological Science 40 (2013) 4576e45854580

sample periods. The coastal samples plotted in Fig. 2 showed anegative linear correlation between d15N and d13C (r ¼ �0.72),whereas valley sites showed a weak positive linear correlationbetween d15N and d13C (r ¼ 0.52). The d15N values for the earliersites from the Archaic to Formative period (2300 BCeAD 500)produced a mean of þ26.5 (1s ¼ 5.6&) compared to later periodsites (Middle Horizon and Late Intermediate period, AD 500-AD1476), which had a mean of þ22.8& (1s ¼ 2.6&). Although the t-test was not statistically-significant, the effect size statistic indi-cated a large difference between the means (t ¼ 1.90, df ¼ 10,p ¼ .09, d ¼ 0.85). The d13C mean was �18.4& (1s ¼ 0.9&) for theearlier sites from the Archaic to Formative period and �17.3&(1s¼ 2.0&) for the later period from theMiddle Horizon to the LateIntermediate period. The independent samples t-test indicated thatthe mean d13C difference between early (Archaic to Formative) andlate (Middle Horizon to Late Intermediate) periods approachedsignificance (t ¼ 1.95, df ¼ 24, p ¼ .06, d ¼ 0.72).

Multiple calculus samples were obtained from seven individuals(see paired, shaded samples in Table 1). For five individuals, isotopesignatures could be compared between the two jaws. In four cases,results could be compared between anterior and posterior teeth.Themean difference in isotope compositions betweenmaxillae andmandibles for the same individual was 0.7& for d15N (1s ¼ 0.4&)and 0.6& for d13C (1s¼ 0.3&). The contrast between anterior teethand posterior teeth for the same individual was 1.3& for d15N(1s ¼ 0.9&) and 0.5& for d13C (1s ¼ 0.3&).

5. Discussion

5.1. Multiple analyses from the same individual

In general, calculus sampled from different teeth for the sameindividual showed small differences in d15N and d13C values. Thedifferences that exist may be explained by the observation thatcalculus accumulates at variable rates in different parts of themouth, so calculus deposited on different teeth may be non-contemporaneous. Thus, calculus could potentially record isotopiccompositions that reflect changes in diet vs. time. However, thenumber of samples analyzed for this test was small, and furtherinvestigation of the homogeneity of dental calculus isotopiccomposition within an individual is warranted.

5.2. Elemental concentrations of calculus

Elemental concentrations of carbon and nitrogen, and the C/Nweight ratios for the prehistoric Chilean samples were higher thanthe concentrations measured for medieval and post-medievalskeletons from Spain (Scott and Poulson, 2012). For the Chileansamples, weight % C ranged from 5.97 to 37.20% (mean of 11.74%, 1sof 5.50%), weight % N ranged from 1.05 to 2.42% (mean of 1.41%, 1sof 0.26%), and C/N weight ratios ranged from 6.1 to 15.4 (mean of8.1, 1s of 2.1). In comparison, weight % C ranged from 1.74 to 8.49%(mean of 4.90%, 1s of 1.06%), weight % N ranged from 0.25 to 1.38%

S.R. Poulson et al. / Journal of Archaeological Science 40 (2013) 4576e4585 4581

(mean of 0.74%, 1s of 0.21%), and C/N weight ratios ranged from 4.1to 10.4 (mean of 6.5, 1s of 1.3) for the Spanish samples (Scott andPoulson, 2012). The explanation for the differences between theChilean samples and the Spanish samples is not clear, but dietarydifferences between the two populations may be a contributoryfactor. Mean values of C/N ratios are similar for the coastal(7.8 � 1.1) vs. valley (8.2 � 2.4) sites in Chile.

The relatively high mean C/N ratios for the two sample sets(Chile 8.1; Spain 6.5), indicates the presence of carbonaceouscompounds other than just protein, as collagen and keratin have atypical C/N ratio of 2.9e3.6, and hair has a typical C/N ratio of 2.9e3.8 (DeNiro, 1985; O’Connell and Hedges, 1999). The C/N ratio forcalculus suggests that protein may comprise part, but not all, of thecarbonaceous fraction, and that carbohydrates and lipids (i.e.carbon-rich compounds) may also contribute to the observed C/Nratios. Because this study is one of the first of its kind to employcalculus as a proxy for isotope analyses, using C/N ratios of calculusfor quality control to identify post-depositional alteration of isotopecompositions is not possible at this time and future research in thisarea is needed. Although limited information concerning theoccurrence of carbonaceous and nitrogenous compounds in dentalcalculus is available (e.g. Lieverse, 1999), a complete accounting of Cand N budgets in dental calculus is necessary. This will allow forsubsequent assessment of the susceptibility of each carbon ornitrogen-bearing compound in calculus to post-depositionaldegradation or alteration, and consequently an improved under-standing of possible quality control criteria that may be applicableto the analysis of dental calculus.

5.3. Nitrogen isotopic composition of Chilean dental calculus

The most notable feature of the isotope results for dental cal-culus is the heavy d15N compositions, with a mean d15N of þ30.4&for samples from coastal sites and þ22.3& for the valley sites.These values were in marked contrast to the mean value of þ11.8&for Spanish dental calculus (Scott and Poulson, 2012), and for someother studies that compare coastal vs. inland sites elsewhere inChile using collagen and other isotope proxies. Studies in centralChile have measured a maximum value for d15N of þ14.6& at acoastal site, and lighter d15N values at inland sites (Falabella et al.,2007, 2008; Sanhueza and Falabella, 2010; Tykot et al., 2009).However, although d15N values for human biomaterials heavierthan þ30& are unusual, they are not without precedent.Aufderheide et al. (1993) measured a single d15N value of þ27.2&for mummified tissue in the Acha ravine, located between thecoastal Arica and Azapa valley sites investigated in this study.Aufderheide et al. (1993) also measured d15N values between þ31.7and þ32.1& in mummified human hair and dura mater at anothercoastal site in Chile (approximately 100 km south of Arica), whileAufderheide et al. (1994) measured d15N values as heavy asþ33.1&in mummified human muscle, tissue, and hair at coastal sites atPisagua, approximately 200 km south of Arica and dating toapproximately 1000 BC, and with similar site provenance to thesamples measured in this study. However, Aufderheide et al. (1993,1994) attributed these heavy d15N values to diagenesis even thoughhair samples were thought to be resistant to diagenesis in theextreme arid conditions of this region.

A possible explanation for heavy d15N values is a diet rich inmarine resources, asmarine dietary items typically have heavy d15Ncompositions. For example, populations with a marine-rich dietsuch as Neo- and Paleo-Aleuts in the north Pacific have mean d15Ncompositions of þ20.3 andþ19.5&, respectively (Byers et al., 2011;Fig. 2), while modern Greenlandic Inuit have a d15N range of þ12.2to þ19.1& (mean of þ16.0&; Buchardt et al., 2007), and a singlesample of Alaskan Inuit dental calculus had a d15N value of þ17.5&

(Scott and Poulson, 2012). A survey of fish in the south Pacific showsa d15N range of þ11.8 to þ22.4& (mean of þ19.0&; Tieszen andChapman, 1992). The magnitude of the isotopic shift between ni-trogen retained in dental calculus and nitrogen in the diet iscurrently unknown, but assuming a nitrogen isotope trophicenrichment of approximately þ3& (e.g. Caut et al., 2009; DeNiroand Epstein, 1981), a diet focusing on fish would produce humanbiomaterials with a d15N range ofþ14.8 to þ25.4&. This d15N rangecorresponds closely to results from the Azapa Valley sites (d15Nrange ofþ17.8 toþ28.5&, mean ofþ22.3&), as well as results fromcoastal sites in the Osmore Valley, Peru (Roca Verde, Wawakiki, andSan Geronimo), with a d15N range of þ16.7 to þ25.0& (meanof þ22.0&; Tomczak, 2001, 2003) and results from a coastal site innorthern Chile (Caleta Vitor) with a d15N mean of þ23.0&, and arange betweenþ17.3 andþ26.7& (Roberts et al., 2013). However, adiet consisting exclusively of south Pacific fish cannot explain thed15NArica coastal samples that have d15Nvalues as heavyasþ33.1&.

Dietary items that could potentially produce the heavy d15Nvalues for calculus (this study) and other biomaterials (Aufderheideet al., 1993, 1994) in northern Chile are isotopically heavy plants.Plant d15N compositions as heavy asþ46& have been measured fornon-carbonized samples dating between 2000 BC and AD 1400from desert locations along the central coast of Peru (DeNiro andHastorf, 1985; Fig. 2). The extremely heavy d15N compositionswere explained by DeNiro and Hastorf (1985) as post-burialdiagenetic alteration, although the samples were in an excellentstate of preservation. Moreover, a suite of laboratory chemical and/or thermal treatments could not experimentally produce the hy-pothesized diagenetic enrichment of d15N compositions (DeNiroand Hastorf, 1985). However, isotopically heavy d15N composi-tions for plant materials (ranging from þ27.9 to þ34.8&) have alsobeen measured in grave goods at Pisagua, Chile (Aufderheide et al.,1994; Fig. 2), although these d15N values were also interpreted to bediagenetically altered, following the conclusion of DeNiro andHastorf (1985).

Elevated d15N values are known to be produced in arid envi-ronments, and a negative correlation between d15N and meanannual rainfall (MAR) has been well documented for plants (e.g.Handley et al., 1999; Heaton, 1987), animals (e.g. Ambrose, 1991;Gröcke et al., 1997; Heaton, 1987; Schwarcz et al., 1999; Sealyet al., 1987) and humans (e.g. Heaton et al., 1986; Schwarcz et al.,1999). For human biomaterial, elevated d15N values are producedby twomechanisms: firstly, the consumption of plant material withelevated d15N values that is produced in arid environments; andsecondly, the excretion of d15N-depleted urea in high concentra-tions as a physiological mechanism to compensate for water stress(Schwarcz et al., 1999). The hyper-aridity of the Arica region of theAtacamawith MAR of 0.5e0.6 mm/year (Clarke, 2006) suggests thepossibility that the elevated d15N values for dietary items producedin such an environment may explain the very heavy d15N valuesmeasured for dental calculus. However, while values of d15N forplant or human material in hyper-arid environments are elevated,they do not appear to exceed d15N values of þ20&. For example,modern plant material from the Namib coast (MAR ¼ 10e20 mm/year) has a mean d15N ¼ þ13.5& (1s ¼ 3.3&) and a maximumvalue of þ18.3& (Heaton et al., 1986), while archaeological plantmaterial from the Sahara at Dakhleh, Egypt (MAR ¼ 0.7 mm/year)has a maximum d15N value of þ18.8& (Schwarcz et al., 1999).Similarly, adult human collagen samples from the Dakhleh sitehave amean d15N value ofþ17.6&, with amaximumvalue ofþ20&(Schwarcz et al., 1999). While it appears that the extreme aridity atArica may contribute towards the production of heavy d15N values,the magnitude of this shift towards heavy d15N values is likelyinsufficient to produce the extremely heavy d15N values measuredin this study (i.e. d15N values heavier than þ30&).

S.R. Poulson et al. / Journal of Archaeological Science 40 (2013) 4576e45854582

A potential alternate mechanism to produce plants withextremely heavy d15N compositions is the acquisition of plant ni-trogen from guano, which is the accumulation of seabird excrementin arid environments. The d15N composition of seabird guano hasextremely heavy d15N compositions, as high as þ50& shown inFig. 2 (also see Table 1 of Szpak et al., 2012a). The phenomenon ispartially due to the ingestion of isotopically-heavy marine re-sources by seabirds. This leads to the production of isotopically-heavy feces, which undergoes post-depositional isotopic enrich-ment due to the loss of isotopically-light nitrogen by ammoniavolatilization (e.g. Li et al., 2012), producing soils with d15N com-positions as heavy as þ43& (Mizota, 2009a, 2009b).

The production of plants with extremely heavy d15N composi-tions has been demonstrated by Szpak et al. (2012a, 2012b), whoperformed modern cultivation experiments to grow maize (Zeamays), using Peruvian seabird guano (d15N ¼ þ26.7&) as a fertil-izer. In these studies, maize d15N compositions as heavy as þ45&were produced (i.e. heavier than the applied guano: Fig. 2). Thisillustrated the control of the applied guano d15N composition, aswell as post-application processes that led to isotopic enrichmentof residual nitrogen-bearing compounds (e.g. ammonia volatili-zation) upon the final, measured d15N composition of maize.Although the modern guano-fertilization experiments were onlyperformed using maize cultivates (Szpak et al., 2012a, 2012b), itseems likely that guano would have a similar impact on other wildand domestic plants. Hence, it is plausible that plants withextremely heavy d15N compositions measured by Aufderheideet al. (1994) and DeNiro and Hastorf (1985), and the hair anddura mater samples with extremely heavy d15N compositionsmeasured by Aufderheide et al. (1993, 1994), do not representdiagenetic alteration. Instead, such results may reflect plants thatwere nourished by isotopically-heavy guano, and human bio-materials that represent a diet that included isotopically-heavy,guano-fertilized plants.

Notably, the coastal sites in northern Chile and Peru whereisotopically heavy plant and human biomaterial samples werecollected are close to ancient guano deposits. This is also the casefor Arica where the present study was conducted. Thus, it isconceivable that guano was used as a fertilizer for plant cultivationduring the time periods studied. Alternatively, non-cultivatedplants located close to guano deposits may have receivednaturally-mediated fertilization due to local emission and re-deposition of nitrogen-bearing compounds, as the d15N composi-tions of modern plants associated with guano-rich seabird coloniescan shift toward d15N compositions as heavy as þ37& (Mizota andNaikatini, 2007).

5.4. Combined variation of nitrogen and carbon isotopiccompositions

d13C values for dental calculus range from�21.8 to�14.6&, withmeans of �18.8� 1.0& vs. �17.4 � 1.7& for coastal and valley sites,respectively (Fig. 2, Table 1). d13C values for human biomaterials arecontrolled by the isotopic composition of dietary components, therelative proportion of the various dietary components, and by thecarbon isotope trophic enrichment effect. The magnitude of theisotope trophic enrichment effect (i.e. the difference in isotopiccomposition of the proxy material and the diet) is dependent uponthe specific identity of the proxy material (e.g. bone, hair, muscletissue, tooth enamel, etc.). The size of the isotope trophic enrich-ment factor is currently unknown for carbon incorporated intodental calculus, but as tropic enrichment factors typically have avalue zþ1& for carbon (e.g. Caut et al., 2009; DeNiro and Epstein,1978), for discussion purposes it is assumed that this value can beapplied when interpreting dental calculus d13C values.

For the Chilean sites sampled in this study, it is possible therewere three major dietary components: plants with a C3 photo-synthetic pathway, which have a mean d13C value of �27 to �26&(e.g. Deines, 1980); plants with a C4 photosynthetic pathway (suchas maize), which have a mean d13C value of �13 to �12& (e.g.Deines, 1980); and marine resources, which vary by trophic leveland range up to approximately �11& (e.g. south Pacific fish:Tieszen and Chapman, 1992). A marine-rich diet for Neo and Paleo-Aleuts produced d13C values of �12.5& (Byers et al., 2011; Fig. 2)while samples from coastal sites of the Osmore Valley, Peru hadd13C values that varied between �13 and �10& (Tomczak, 2001,2003; Fig. 2).

The d15N composition of marine resources has been discussed inthe preceding section. Aside from nitrogen fixing plants (e.g. le-gumes), the typical d15N compositions of C3 and C4 plants are notwell defined, as d15N compositions are, in large part, controlled bythe d15N composition of the nitrogen compounds available for useby the plants. Most plants have d15N compositions in the rangeof �5 to þ4& (e.g. Fry, 1991; Horn et al., 2009). Hence, diets con-sisting of varying proportions of two components, such as marineand terrestrial C3 diet components, can produce a positive corre-lation between d15N and d13C values. The dietary isotope mixingline, where the d15N and d13C compositions are controlled by theisotopic compositions of the two dietary components, reflects therelative proportion of the two components in the diet. This resultsin an array of isotopic compositions that fall along a straight line. Anexample of such a dietary isotope mixing line is shown in Fig. 2,with results for prehistoric groups from the Santa Barbara ChannelIslands and mainland of central California, an area that neveradopted maize cultivation or agriculture, with d15N values thatranged from þ7.5 to þ18.7& (Walker and DeNiro, 1986). Similarly,samples from the Tierra del Fuego region of southern Chile show awide range of isotopic compositions, with d15N values from þ9.9to þ18.8& (Yesner et al., 2003).

Calculus from the coastal samples show a moderate negativelinear correlation between d15N and d13C. Previous analyses of hairfrom Arica (Macko et al., 1999) fell along this same trend, but withsomewhat lighter d15N values and somewhat heavier d13C valuesthan those reported in this study. The mummified samples re-ported by Aufderheide et al. (1994) from coastal Pisagua showed aconsiderable overlap of d15N and d13C values with the coastalsamples from this study (Fig. 2). In contrast, the Azapa Valley sitesshowed a moderate positive linear correlation between d15N andd13C derived from dental calculus. The positive linear correlationbetween d15N and d13C for the Azapa Valley samples can potentiallybe explained by a diet that relied on fish and other marine re-sources. Given the assumption that south Pacific fish in the past hadsimilar isotopic compositions to extant fish, and assuming a trophicshift of þ3& for d15N and þ1& for d13C, the positive correlationobserved between d15N and d13C for the fish compositions thatmade up the diet (Tieszen and Chapman, 1992; Fig. 2) would alsoproduce a positive correlation between d15N and d13C for humanbiomaterials. In this case, there would be a parallel correlation lineshifted by approximately þ3& for d15N vs. the fish isotopiccomposition correlation line. Moreover, Aufderheide and Santoro(1999) reached a similar conclusion in the Lluta Valley, a few ki-lometers fromArica and the Azapa Valley; despite the abundance ofmacrobotanical remains from archaeological excavations, chemicalanalyses showed that fish and shellfish were a significant dietarycomponent for late prehistoric groups.

It is impossible to produce the negative linear correlation be-tween d15N and d13C observed for the Arica coastal samples usingan exclusively marine-derived diet, nor is it possible to produced15N values for biomaterials heavier thanþ30& (Aufderheide et al.,1993, 1994, this study; Fig. 2), as the heaviest d15N composition for

S.R. Poulson et al. / Journal of Archaeological Science 40 (2013) 4576e4585 4583

fish is þ22.4& (Tieszen and Chapman, 1992). A possible explana-tion for d15N values heavier than þ30& and the negative linearcorrelation between d15N and d13C could be a two-component dietconsisting of a marine component (d15N z þ22&, d13C z �12&),with a C3 plant component that has been nourished by isotopicallyheavy guano (d15N z þ40&, d13C z �26&). Indeed, plants withsimilar d15N and d13C compositions have been measured for sam-ples from coastal sites in Peru (DeNiro and Hastorf, 1985; Fig. 2), butC3 plant samples with these compositions have not been measuredfor any sites near Arica or Pisagua. However, compelling evidenceobtained through coprolite samples from coastal Archaic-periodArica mummies demonstrated that plant foods were indeed pre-sent and part of a broad-spectrum diet that included marine re-sources and wild plants (Reinhard et al., 2011). There does notappear to be any direct archaeological evidence for extensivecultivation of C3 plants, rather than C4 plants such as maize.Nevertheless, it is difficult to envision an alternative hypothesis toexplain d15N values heavier than þ30& and a negative linear cor-relation between d15N and d13C.

5.5. Archaeological and historic evidence for the use of guano in theAndes

In the nineteenth century, seabird guano deposits were minedfrom Chile and Peru for export as a fertilizer to Europe and NorthAmerica. Guano from these deposits, in high demand because of itshigh nitrogen concentration, was an important resource until thedevelopment of synthesized chemical fertilizers in the earlytwentieth century (Skaggs, 1994). There is little direct archaeolog-ical evidence for the prehispanic use of crop fertilizers, like seabirdand camelid guano, in northern Chile and other regions of theAndes. However, some circumstantial evidence for prehispanicfertilizers exists. Szpak et al. (2012b) noted that prehistoric ceramicartifacts have been found within guano deposits off the shores ofthe north coast of Peru. Rivera (1991) indicated that the Early In-termediate period (500 BCeAD 500) of cultural development innorthern Chile was notable for the production of surplus and tradegoods in the Pacific coastal valleys, including items such as maize,peppers, potatoes, and guano. Historical sources also have revealedthe importance of guano to local people in northern Chile. The useof guano as a crop fertilizer was documented by sixteenth andseventeenth century Spanish chronicler, Cieza de León (2005[1553]: 205), who indicated that the people of the Tarapaca Val-ley in northern Chile were not only familiar with the importance ofseabird guano for growing corn, but that they traveled to the coastand sailed out to the rocky islands just off the shore to collect thefertilizer. Frézier (1902[1714]) documented similar accounts for thepeople living in the Bay of Arica, in which he noted that the localgrowers added a small handful of guano to the bottom of eachpepper plant.

Historic accounts from the Spanish chroniclers do not providedirect evidence for the use of seabird guano as a fertilizer duringprehistoric times. Although genetic data have demonstrated directevidence of plant domestication in the New World by at least 8000years BP (see Erickson et al., 2005), there is limited direct archaeo-logical evidence for crop cultivation before the Formative period atcoastal sites in the extreme north of Chile. Pearsall (2008:112) notedthat maize was recovered at Archaic-period sites of Tilliviche 1-b(5900e4110 BC), Camarones 14 (5470e4665 BC), and the QuianiShellmound (4220e3680 BC) on the coast. Archaeological excava-tions at the Quiani site, also located on the north coast of Chile,producedmaize, gourd, cotton and quinoa at levels dating to around2500 BC (Pearsall, 2008). However, archaeologists working in theregion have questioned such finds, indicating that maize and othercultivates may have been associated with later strata in the

shellmounds. The first archaeological evidence of maize cultivationin the Arica Valleys dates to w1000 BC (Aufderheide et al., 2002;Muñoz Ovalle, 2004). The Archaic and later sites sampled in thisstudy were indeed in close contact with seabirds and seabird guanobecause their plumes, beaks, and bones have been recovered inarchaeological middens and burial contexts (Arriaza et al., 2008;Standen, 2003). Nonetheless, even the consumption of seabirdswould not produce d15N values as heavy as those reported in theextreme north of Chile. Given these lines of evidence, and theabundance of seabirds and guano deposits in Arica and the regionduring ancient times (Llagostera,1979), it is conceivable that peoplemaynothave intentionally usedguano as a fertilizer, but that guano-d15N signatures may have accidentally been incorporated into thediet through local plants that were naturally fertilized by guano.

6. Conclusions

Following Scott and Poulson (2012), stable isotope analysis ofdental calculus was performed to investigate paleodiet for a suite ofsamples from a desert region of northern Chile. Calculus wascollected from the teeth of individuals excavated from seven coastand Azapa Valley archaeological deposits in Arica with a broadtemporal sequence, from the Archaic (w2300 BC) through the LateIntermediate period (AD 1476). Samples from the Azapa Valley sitesyielded isotopically heavy d15N compositions (range of þ17.8to þ28.5&), and showed a positive correlation between d15N andd13C. The hyper-aridity at the study site (mean annual rainfall of 0.5e0.6mm/year), canhelp explain someof the shift ofd15Ncompositionstowards heavy values, but is unlikely to account for the presence ofthe heaviest d15N values (i.e. d15N > þ20&). Similarly, some of theresults from the Azapa Valley sites can potentially be explained by amarine-rich diet, a conclusion also reached through recent bio-archaeological and isotope analyses (e.g. Roberts et al., 2013;Watsonet al., 2011), or possibly a marine-heavy diet supplemented withterrestrial meat and plant resources, as suggested by trace elementdata (Aufderheide and Allison, 1992; as cited in Arriaza, 1995b).Notably, these data coincide with archaeological evidence for theemergence of agriculture and crop cultivation in the region (MuñozOvalle, 2004; Pearsall, 2008) and with analyses conducted in theadjacent Lluta Valley (Aufderheide and Santoro, 1999).

Samples from the Arica coastal sites yielded extremely heavyd15N compositions (ranging from þ26.2 to þ33.1&), and showed anegative correlation between d15N and d13C. This was consistentwith analyses of other biomaterials (nail, muscle, hair, dura mater)measured elsewhere in coastal Chile (Aufderheide et al.,1993,1994).These extremely heavy d15N values cannot be produced by thehyper-arid environment and/or a marine-rich diet. A suggestedexplanation for these extremely heavy d15N values is a two-component diet consisting of a marine component (d15N z þ22&,d13C z �12&), with a C3 plant component that has been treatedwith isotopically-heavy guano (d15N z þ40&, d13C z �26&).Supporting evidence for this explanation includes: (1) analyses ofplants collected from other coastal archaeological sites in Chile andPeruwith d15N compositions heavier thanþ35& (Aufderheide et al.,1994; DeNiro and Hastorf, 1985); (2) modern plant cultivation ex-periments with guano fertilizer that can produce plants with d15Ncompositions heavier than þ40& (Szpak et al., 2012a, 2012b); and(3) coprolite analyses that showed native plant foods were animportant dietary component during the Archaic period (Reinhardet al., 2011). Possible confounding evidence for this explanationincludes: (1) the absence of any direct archaeological evidence forcrop cultivation (especially for C3 plants) before the Formativeperiod at coastal sites in northern Chile; and (2) the lack of any directevidence for the use of seabird guano as a fertilizer during prehis-toric times, although circumstantial evidence suggests this may

S.R. Poulson et al. / Journal of Archaeological Science 40 (2013) 4576e45854584

have occurred. Nevertheless, the use of seabird guano fertilizer toproduce an extremely heavy d15N signature in human biomaterialsis an intriguing area of research that warrants further study incoastal areas with extensive guano deposits, such as Chile and Peru.

In prehistoric populations, calculus would accumulate from thetime of initial tooth eruption to the death of an individual. Calculuscan absorb a variety of elements that reside in or pass through theoral cavity, but in contrast to the smooth enamel surface of a toothcrown, the external face of calculus is rough and porous, making itsusceptible to further plaque and calculus formation. However,calculus is not the same as collagen or keratin, and additionalresearch is required to determine if stable carbon and nitrogenisotope compositions preserved in calculus are measuring compa-rable or different dietary inputs to the isotopic compositions pre-served in collagen or keratin. Despite this caveat, dental calculusmay be an effective and complementary proxy for obtaining reli-able isotope signatures for paleodietary research. Additionally,calculus may serve as a biomaterial that museum and collectioncurators are more willing to provide for destructive analysis, due toits nature as a secondary biomaterial (or “add-on”).

Acknowledgments

The authors would like to extend their gratitude to HectorGonzalez Cortez at the Universidad de Tarapacá and Museo SanMiguel de Azapa for permission to sample the skeletal remainsused in this study and to SusanaMonsalve, who assisted uswith thecollections. This study was funded in part by an IIE FulbrightFellowship (SCK), Project Fondecyt 1121102 (VGS), and the Uni-versity of Nevada, Reno. Finally, the authors thank three anony-mous reviewers whose comments helped us improve the finalversion of this manuscript.

References

Ambrose, S.H., 1991. Effects of diet, climate and physiology on nitrogen isotopeabundances in terrestrial foodwebs. Journal of Archaeological Science 18,293e317.

Arriaza, B.T., 1995a. Chinchorro bioarchaeology: chronology and mummy seriation.Latin American Antiquity 6, 35e55.

Arriaza, B.T., 1995b. Beyond Death. The Chinchorro Mummies of Ancient Chile.Smithsonian Institution Press, Washington.

Arriaza, B.T., Standen, V.G., Cassman, V., Santoro, C.M., 2008. Chinchorro culture:pioneers of the coast of the Atacama Desert. In: Silverman, H., Isbell, W.H. (Eds.),The Handbook of South American Archaeology. Springer, New York, pp. 45e58.

Arriaza, B.T., Reinhard, K.J., Araujo, A.D., Orellana, N.C., Standen, V.G., 2010. Possibleinfluence of the ENSO phenomenon on the pathoecology of diphyllobothriasisand anisakiasis in ancient Chinchorro populations. Memorias do InstitutoOswaldo Cruz 105, 66e72.

Aufderheide, A.C., Allison, M., 1992. Chemical Dietary Reconstruction of North ChilePrehistoric Populations by Trace Mineral Analysis. Manuscript on File. Depart-ment of Pathology, University of Minnesota, Duluth.

Aufderheide, A.C., Muñoz, I., Arriaza, B., 1993. Seven Chinchorro mummies and theprehistory of northern Chile. American Journal of Physical Anthropology 91,189e201.

Aufderheide, A.C., Kelley,M.A., Rivera,M., Gray, L., Tieszen, L.L., Iverson, E., Krouse,H.R.,Caravic, A., 1994. Contributions of chemical dietary reconstruction to the assess-ment of adaptation by ancient highland immigrants (Alto Ramirez) to coastalconditions at Pisagua, north Chile. Journal of Archaeological Science 21, 515e524.

Aufderheide, A.C., Santoro, C.M., 1999. Paleodietary reconstruction: human pop-ulations at late prehistoric sites in the Lluta Valley of northern Chile. RevistaChilena de Historia Natural 72, 237e250.

Aufderheide, A.C., Aturaliya, S., Focacci, G., 2002. Enfermedades pulmonares de unamuestra de población del cementerio AZ-75, Valle de Azapa, norte de Chile.Chungará 34, 253e263.

Bryson, L.M., Bryson, R.U., Bryson, R.A., 2001. Paleoclimatic and material culturalperspective on the formative period of Northern Chile. Chungará 33, 5e12.

Buchardt, B., Bunch, V., Helin, P., 2007. Fingernails and diet: stable isotope signa-tures of a marine hunting community from modern Uummannaq, northGreenland. Chemical Geology 244, 316e329.

Byers, D.A., Yesner, D.R., Broughton, J.M., Coltrain, J.B., 2011. Stable isotope chem-istry, population histories and Late Prehistoric subsistence change in theAleutian Islands. Journal of Archaeological Science 38, 183e196.

Caut, S., Angulo, E., Courchamp, F., 2009. Variation in discrimination factors d15Nand d13C: the effect of diet isotopic values and applications for diet recon-struction. Journal of Applied Ecology 46, 443e453.

Cieza de León, P., 2005. Crónica del Perú el Señorío de los Incas [1553]. FundaciónBiblioteca, Caracas.

Clarke, J.D.A., 2006. Antiquity of aridity in the Chilean Atacama Desert. Geo-morphology 73, 101e114.

Dauelsberg, P., 1974. Excavaciones arqueológicas en Quiani. Chungará 4, 7e38.Deines, P., 1980. The isotopic composition of reduced carbon. In: Fritz, P., Fontes, J.C.

(Eds.), Handbook of Environmental Isotope Geochemistry, The TerrestrialEnvironment, vol. 1. Elsevier, Amsterdam, pp. 329e406.

DeNiro, M.J., 1985. Postmortem preservation and alteration of in vivo bone collagenisotope ratios in relation to palaeodietary reconstruction. Nature 317, 806e809.

DeNiro, M.J., Epstein, S., 1978. Influence of diet on the distribution of carbon iso-topes in animals. Geochimica et Cosmochimica Acta 42, 495e506.

DeNiro, M.J., Epstein, S., 1981. Influence of diet on the distribution of nitrogenisotopes in animals. Geochimica et Cosmochimica Acta 45, 341e351.

DeNiro, M.J., Hastorf, C.A., 1985. Alteration of 15N/14N and 13C/12C ratios of plantmatter during the initial stages of diagenesis: studies utilizing archaeologicalspecimens from Peru. Geochimica et Cosmochimica Acta 49, 97e115.

Dillehay, T.D., 2008. Profiles in Pleistocene history. In: Silverman, H., Isbell, W.H.(Eds.), The Handbook of South American Archaeology. Springer, New York,pp. 29e43.

Erickson, D.L., Smith, B.D., Clarke, A.C., Sandweiss, D.H., Tuross, N., 2005. An Asianorigin for a 10,000-year-old domesticated plant in the Americas. Proceedings ofthe National Academy of Sciences U. S. A. 102, 18315e18320.

Falabella, F., Planella, M.T., Aspillaga, E., Sanhueza, L., Tykot, R.H., 2007. Dieta enSociedades Alfareras de Chile Central: Aporte de Análisis de Isótopos Estables.Chungará 39, 5e27.

Falabella, F., Planella, M.T., Tykot, R.H., 2008. El Maíz (Zea Mays) en el Mundo Pre-hispánico de Chile Central. Latin American Antiquity 19, 25e46.

Focacci, G., 1974. Excavaciones en Playa Miller 7, Arica. Chile. Chungará 3, 23e24.Focacci, G., 1990. Excavaciones arqueológicas en el cementerio AZ-6 Valle de Azapa.

Chungará 24/25, 69e124.Focacci, G., Chacon, S., 1989. Excavaciones arqueológicas en los faldeos del morro de

Arica, sitios Morro-1/6 y 2/2. Chungará 22, 25e62.Frézier, A.F., 1902. Relación del Viaje por el Mar del Sur a las costas de el Perú y el

Chile [1714]. Imprenta Mejia, Santiago.Fry, B.,1991. Stable isotope diagramsof freshwater foodwebs. Ecology 72, 2293e2297.Gröcke, D.R., Bocherens, H., Mariotti, A., 1997. Annual rainfall and nitrogen-isotope

correlation in macropod collagen: application as a paleoprecipitation indicator.Earth and Planetary Science Letters 153, 279e285.

Handley, L.L., Austin, A.T., Robinson, D., Scrimgeour, C.M., Raven, J.A., Heaton, T.H.E.,Schmidt, S., Stewart, G.R., 1999. The 15N natural abundance (d15N) of ecosystemsamples reflects measures of water availability. Australian Journal of PlantPhysiology 26, 185e199.

Hayashizaki, J., Ban, S., Nakagaki, H., Okumura, A., Yoshii, S., Robinson, C., 2008. Sitespecific mineral composition and microstructure of human supra-gingivaldental calculus. Archives of Oral Biology 53, 168e174.

Heaton, T.H.E., Vogel, J.C., von la Chevallerie, G., Collett, G., 1986. Climatic influenceon the isotopic composition of bone nitrogen. Nature 322, 822e823.

Heaton, T.H.E., 1987. The 15N/14N ratios of plants in South Africa and Namibia: rela-tionship to climates and coastal/saline environments. Oecologia 74, 236e246.

Horn, P., Hölzl, S., Rummel, S., Åberg, G., Schiegl, S., Daniela Biermann, D., Struck, U.,Rossmann, A., 2009.Humans and camelids inRiverOases of the IcaePalpaeNazcaregion inpre-hispanic timese insights fromH-C-N-O-S-Sr isotope signatures. In:Reindel, M., Wagner, G.A. (Eds.), New Technologies for Archaeology: Multidisci-plinary Investigations in Palpa and Nasca, Peru. Springer, Berlin, pp. 173e192.

Kelley, M.A., Levesque, D.r., Weidl, E., 1991. Contrasting patterns of dental disease infive early northern Chilean groups. In: Kelley, M.A., Larson, C.S. (Eds.), Advancesin Dental Anthropology. Wiley-Liss, New York, pp. 203e213.

Li, L., Lollar, B.S., Li, H., Wortmann, U.G., Lacrampe-Couloume, G., 2012. Ammoniumstability and nitrogen isotope fractionations for NH4

þeNH3(aq)eNH3(gas) systemsat 20e70 �C and pH of 2e13: applications to habitability and nitrogen cycling inlow-temperature hydrothermal systems. Geochimica et Cosmochimica Acta 84,280e296.

Lieverse, A.R., 1999. Diet and the aetiology of dental calculus. International Journalof Osteoarchaeology 9, 219e232.

Llagostera, A.,1979. 9,700 years of maritime subsistence on the Pacific: an analysis bymeans of bioindicators in the north of Chile. American Antiquity 44, 309e324.

Macko, S.A., Engel, M.H., Andrusevich, V., Lubec, G., O’Connell, T.C., Hedges, R.E.M.,1999. Documenting the diet of ancient human populations through stableisotope analysis of hair. Philosophical Transactions of the Royal Society B:Biological Sciences 354, 65e76.

Marquet, P.A., Santoro, C.M., Latorre, C., Standen, V.G., Abades, S.n.R.,Rivadeneira, M.M., Arriaza, B., Hochberg, M.E., 2012. Emergence of socialcomplexity among coastal hunteregatherers in the Atacama Desert of northernChile. Proceedings of the National Academy of Sciences 109, 14754e14760.

Mizota, C., 2009a. Temporal variations in the concentration and isotopic signatureof ammonium- and nitrate-nitrogen in soils under a breeding colony of Black-tailed Gulls (Larus crassirostris) on Kabushima Island, northeastern Japan.Applied Geochemistry 24, 328e332.

Mizota, C., 2009b. Nitrogen isotopic patterns of vegetation as affected by breedingactivity of Black-tailed Gull (Larus crassirostris): a coupled analysis of feces,inorganic soil nitrogen and flora. Applied Geochemistry 24, 227e2033.

S.R. Poulson et al. / Journal of Archaeological Science 40 (2013) 4576e4585 4585

Mizota, C., Naikatini, A., 2007. Nitrogen isotope composition of inorganic soil ni-trogen and associated vegetation under a sea bird colony on the Hatana islands,Rotuma Group, Fiji. Geochemical Journal 41, 297e301.

Muñoz Ovalle, I., 2004. El Periodo formativo en los valles del norte de Chile y sur dePerú: nuevas evidencias y comentarios. Chungará Special Volume, 213e225.

Muñoz Ovalle, I., Focacci, G.A., 1985. San Lorenzo: Testimonio de una comunidad deagricultores y pescadores Postiwanaku en al valle de Azapa (Arica-Chile).Chungará 15, 7e30.

O’Connell, T.C., Hedges, R.E.M., 1999. Investigations into the effect of diet on modernhuman hair isotopic values. American Journal of Physical Anthropology 108,409e425.

Pearsall, D.M., 2008. Plant domestication and the shift to agriculture in the Andes.In: Silverman, H., Isbell, W.H. (Eds.), The Handbook of South AmericanArchaeology. Springer, New York, pp. 105e120.

Reinhard, K., Urban, O., 2003. Diagnosing ancient diphyllobothriasis from Chin-chorro mummies. Memórias do Instituto Oswaldo Cruz, Rio de Janeiro 98(Suppl. I), 191e193.

Reinhard, K.J., LeRoy-Toren, S., Arriaza, B., 2011. Where have all the plant foodsgone? The search for refined dietary reconstruction from Chinchorro mummies.Yearbook of Mummy Studies 1, 139e151.

Rivera, M.A., 1991. Prehistory of northern Chile: a synthesis. Journal of World Pre-history 5, 1e47.

Rivera, M.A., 2008. The archaeology of northern Chile. In: Silverman, H., Isbell, W.H.(Eds.), The Handbook of South American Archaeology. Springer, New York,pp. 963e977.

Roberts, A., Pate, F.D., Petruzzelli, B., Carter, C., Westaway, M.C., Santoro, C.M., Swift, J.,Maddern, T., Jacobsen,G.E., Bertuch, F., Rothhammer, F., 2013.Retentionofhunteregatherer economies among maritime foragers from Caleta Vitor, northern Chile,during the late Holocene: evidence from stable carbon and nitrogen isotopic an-alyses of skeletal remains. Journal of Archaeological Science 40, 2360e2372.

Rothhammer, F., Santoro, C.M., 2001. El Desarollo cultural en el valle de Azapa,extremo norte de Chile y su vinculación con los desplazamientos poblacionalesaltiplánicos. Latin American Antiquity 12, 59e66.

Sanhueza, L., Falabella, F., 2010. Analysis of stable isotopes: from the archaic to thehorticultural communities in central Chile. Current Anthropology 51, 127e136.

Santoro, C., Dorsey Vinton, S., Reinhard, K.J., 2003. Inca expansion and parasitism inthe Lluta Valley: preliminary data. Chungará 98, 161e163.

Schoeninger, M.J., 2009. Stable isotope evidence for the adoption of maize agri-culture. Current Anthropology 50, 633e640.

Schoeninger, M.J., 2010. Diet reconstruction and ecology using stable isotope ratios.In: Larsen, C.S. (Ed.), A Companion to Biological Anthropology. Wiley-Blackwell,Chichester, U.K., pp. 445e464.

Schwarcz, H.P., Dupras, T., Fairgrieve, S.I., 1999. 15N enrichment in the Sahara: insearch of a global relationship. Journal of Archaeological Science 26, 629e636.

Schwarcz, H.P., Schoeninger, M.J., 1991. Stable isotope analyses in human nutritionalecology. Yearbook of Physical Anthropology 34, 283e321.

Scott, G.R., Poulson, S.R., 2012. Stable carbon and nitrogen isotopes of human dentalcalculus: a potentially new non-destructive proxy for paleodietary analysis.Journal of Archaeological Science 39, 1388e1393.

Sealy, J.C., van der Merwe, N.J., Thorp, J.A.L., Lanham, J.L., 1987. Nitrogen isotopicecology in southern Africa: implications for environmental and dietary tracing.Geochimica et Cosmochimica Acta 51, 2707e2717.

Skaggs, J.M., 1994. The Great Guano Rush: Entrepreneurs and American OverseasExpansion. St. Martins Press, New York, p. 334.

Standen, V.G., 2003. Funerary goods from Chinchorro Morro 1 cemetery: descrip-tion, analysis, and interpretation. Chungará 35, 175e207.

Standen, V.G., Arriaza, B.T., Santoro, C.M., 1997. External auditory exostosis in pre-historic Chilean populations: a test of the cold water hypothesis. AmericanJournal of Physical Anthropology 103, 119e129.

Standen, V.G., Santoro, C.M., Arriaza, B.T., 2004. Síntesis y propuestas para elperiodo Arcaico en la costa del extremo norte de Chile. Chungará 36, 201e212.

Sutter, R.C., Mertz, L., 2004. Nonmetric cranial trait variation and prehistoric bio-cultural change in the Azapa Valley, Chile. American Journal of Physical An-thropology 123, 130e145.

Szpak, P., Longstaffe, F.J., Millaire, J.-F., White, C.D., 2012a. Stable isotope biogeo-chemistry of seabird guano fertilization: results from growth chamber studieswith maize. PLoS ONE 7, e33741.

Szpak, P., Millaire, J.F., White, C.D., Longstaffe, F.J., 2012b. Influence of seabird guanoand camelid dung fertilization on the nitrogen isotopic composition of field-grown maize (Zea mays). Journal of Archaeological Science 39, 3721e3740.

Tieszen, L.L., Chapman, M., 1992. Carbon and nitrogen isotopic status of the majormarine and terrestrial resources in the Atacama Desert of northern Chile. In:Proceedings of the First World Congress on Mummy Studies. Museo Arqueo-lógico y Etnográfico de Tenerife, Santa Cruz de Tenerife, pp. 409e425.

Tomczak, P.D., 2001. Prehistoric Socio-Economic, Relations and Population Orga-nization in the Lower Osmore Valley of Southern Peru. Ph.D. dissertation.University of New Mexico, Albuquerque, p. 317.

Tomczak, P.D., 2003. Prehistoric diet and socioeconomic relationships within theOsmore Valley of southern Peru. Journal of Anthropological Archaeology 22,262e278.

Tykot, R.H., Falabella, F., Planella, M.T., Aspillaga, E., Sanhueza, L., Becker, C., 2009.Stable isotopes and archaeology in central Chile: methodological insights andinterpretative problems for dietary reconstruction. International Journal ofOsteoarchaeology 19, 156e170.

Walker, P.L., DeNiro, M.J., 1986. Stable nitrogen and carbon isotope ratios in bonecollagen as indices of prehistoric dietary dependence on marine and terrestrialresources in Southern California. American Journal of Physical Anthropology 71,51e61.

Watson, J.T., Muñoz Ovalle, I., Arriaza, B.T., 2010. Formative adaptations, diet, andoral health in the Azapa Valley of northwest Chile. Latin American Antiquity 21,423e439.

Watson, J.T., Arriaza, B., Standen, V., Muñoz Ovalle, I., 2011. Tooth wear related tomarine foraging, agro-pastoralism and the formative transition on the northernChilean coast. International Journal of Osteoarchaeology. http://dx.doi.org/10.1002/oa.1247.

Werner, R.A., Bruch, B.A., Brand, W.A., 1999. ConFlo III e an interface for highprecision d13C and d15N analysis with an extended dynamic range. RapidCommunications in Mass Spectrometry 13, 1237e1241.

White, D.J., 1997. Dental calculus: recent insights into occurrence, formation, pre-vention, removal and oral health effects of supragingival and subgingival de-posits. European Journal of Oral Sciences 105, 508e522.

Yesner, D.R., Torres, M.J.F., Guichon, R.A., Borrero, L.A., 2003. Stable isotope analysisof human bone and ethnohistoric subsistence patterns in Tierra del Fuego.Journal of Anthropological Archaeology 22, 279e291.