eating fish in the driest desert in the world: osteological and

In the Andes, research on the social nature ofprehistoric trade and interaction has yieldedfoundational works including Murra’s (1972,

1985) vertical archipelago, Núñez and Dillehay’s(1995) movilidad giratoria, and Martínez’s (1998)territorialidad salpicada. While these themes areof ongoing interest, much archaeological researchhas focused on patterns of trade in large, complexsocieties while neglecting small-scale exchange(Berenguer 2004; Dillehay et al. 2006; Nielsen

2004; Schortman and Urban 1992). This narrowvision has had some negative effects in the Andes,where archaeologists have placed great emphasison the exchange networks, llama caravans, andextensive trade routes of large states and empiresat the expense of smaller-scale/intensity trade inter-actions (e.g., Malpass 1993; Trombold 1991).

More recently, archaeologists have stepped awayfrom binary notions of core and periphery, begin-ning to explore supposed empty spaces— areas

EATING FISH IN THE DRIEST DESERT IN THE WORLD: OSTEOLOGICAL AND BIOGEOCHEMICAL ANALYSES OF HUMAN

SKELETAL REMAINS FROM THE SAN SALVADOR CEMETERY,NORTH CHILE

Christina Torres-Rouff, William J. Pestle, and Francisco Gallardo

The San Salvador River in northern Chile is a tributary of the Loa, the only river that traverses the Atacama Desert fromthe mountains to the Pacific. Recent investigations along the San Salvador River revealed the presence of a Formative periodvillage site and associated cemetery approximately 110 km inland from the Pacific Ocean. Bioarchaeological and biogeo-chemical analyses conducted on human skeletal remains recovered from the San Salvador Cemetery allowed us to betterunderstand the site’s role as an intermediary for coast/interior relations during the Middle Formative (500 B.C.–A.D. 100).Evidence from material culture and human remains at San Salvador suggests that this population was involved in exchangenetworks that united the oases of the Atacama Desert with the Pacific Ocean. Isotopic data support this notion, as the pop-ulation demonstrates great variability in both the protein (marine and terrestrial) and carbohydrate components of theirdiet. During this period, communal economies produced surpluses used in a network of exchange for foods, prestige goods,and ideas. These ties were not coincidental but, rather, part of a regional economic structure that remains only partly explored.

El Río San Salvador del norte de Chile es un afluente del Río Loa, el único curso de agua dulce que atraviesa el desierto deAtacama desde las montañas hasta el Pacífico. Recientes investigaciones en San Salvador revelaron la presencia de una aldeay un cementerio del periodo Formativo emplazados a unos 110 km. de la costa del Pacifico. Análisis bioantropológicos y bio-geoquímicos llevados a cabo en restos óseos humanos recuperados del cementerio permitieron comprender el funcionamientodel sitio como intermediario en las relaciones entre la costa y el interior en el Formativo Medio (500 a.C.–100 d.C.). La evi-dencia de la cultura material y de los restos humanos en San Salvador indica que esta población estuvo involucrada en lasredes de intercambio que ligaban los oasis del desierto de Atacama y el Océano Pacífico. Los datos isotópicos apoyan esto,pues la población muestra una gran variabilidad tanto en el consumo de proteína (marina y terrestre) como en los compo-nentes de carbohidratos en su dieta. Durante este período las economías comunales produjeron excedentes utilizados en unared de intercambio de alimentos, bienes de prestigio e ideas. Estos lazos no fueron circunstanciales, sino parte de una estruc-tura económica regional que hasta ahora ha sido parcialmente explorada.

Christina Torres-Rouff ! Instituto de Investigaciones Arqueológicas y Museo, Universidad Católica del Norte, CalleGustavo Le Paige #380, San Pedro de Atacama, 141-0000, Chile; and Department of Anthropology, Colorado College([email protected])William J. Pestle ! Department of Oral Medicine and Diagnostic Sciences (MC838), College of Dentistry Room 569E,University of Illinois, Chicago, 801 S. Paulina, Chicago, IL 60612 ([email protected])Francisco Gallardo ! Museo Chileno de Arte Precolombino, Bandera 361, Santiago, Chile ([email protected])

Latin American Antiquity 23(1), 2012, pp. 51–69Copyright ©2012 by the Society for American Archaeology

51

52 LATIN AMERICAN ANTIQUITY [Vol. 23, No. 1, 2012

between population centers that, while not princi-pal nodes, may have served as loci for differentforms of relationships between groups (e.g., Caseset al. 2008; Nielsen 2004; Upham 1992). Lazzarihas argued that for Formative period northwestArgentina, scholars need “to think of circulationnetworks . . . as more flexible and complicated andas a reflection of the constitution and preproductionof different social bonds” (2005a:202). Considera-tion of varied mobility strategies may provide onemeans of approaching this flexibility. Nielsen(2004:33) stresses the value of internodal spaces forhelping understand the negotiation of social rela-tions by considering the role of individual actors andargues that this type of study requires the use of mul-tiple lines of evidence. Here, by focusing on skele-tal remains from a small cemetery associated witha settlement that likely facilitated some movementsof peoples, goods, and ideas across the Atacama

Desert during the Formative period (1500B.C.–A.D. 500), we complement archaeologicalresearch in the region by providing bioarchaeolog-ical and biogeochemical perspectives on movementand exchange between northern Chile’s coast andinterior. In doing so, we explore the idea that thesevaried movements between population groups andacross the open desert included individuals in tradeoutposts located in otherwise unpopulated areas.The individuals and settlements found therein canhelp to shed light on the way that prehistoric peo-ples used these spaces and the dynamic createdwhen groups interacted. Pimentel and colleagues(2010) posit that in northern Chile’s Formativeperiod there were at least two forms of mobility, onebeing a small-scale pedestrian movement instigatedby coastal populations in search of resources andthe other encompassing the highland caravanmodality. Evidence from this series of excavations

Figure 1. Map of the study area indicating sites mentioned in the text.

suggests that the sites along the San Salvador werelogistically situated occupations resulting fromincreasing social complexity and increasingly com-plex patterns of interaction between the coast andhighlands, processes that had already begun at theend of the Archaic period (Ballester and Gallardo2011; de Souza et al. 2010; Gallardo 2009). As such,the San Salvador sites served as stable and strate-gic outposts for the travelers who connected theMiddle Loa with the coast. This small sample pro-vides us with an interesting view on mobility in theFormative period.

San Salvador

The San Salvador is a tributary of the Loa River(Figure 1), the only river that traverses the breadthof the hyperarid Atacama Desert of northern Chilefrom the mountains to the Pacific. During the For-mative period in this region we see burgeoningnumbers of permanent and semipermanent settle-ments in combination with plant and animal domes-tication, as well as intensified camelid pastoralism(Gallardo 2009; Lumbreras 2006; Núñez, Carta-jena, Carrasco, de Souza, and Grosjean 2006). Pas-toralism and plant collecting form the basis of thisperiod’s economy, despite the beginnings of small-scale agriculture and the intensification of hunting(Agüero 2005; Núñez, Cartajena, Carrasco, and deSouza 2006; Sinclaire 2004). Archaeological evi-dence from the Formative period has demonstrateda flow of goods through this region between moreconsolidated centers, or nodes, such as Tarapacá inthe north and the San Pedro de Atacama oases inthe interior (Agüero et al. 2004). By the Middle For-mative (500 B.C.–A.D. 100), there is evidence ofexchange throughout the region, from the Pacificcoast to northwest Argentina. This accumulation ofsocial and economic changes is likely tied to therise of surplus economies, not only among settledagriculturalists but also for the marine hunter-gath-erers of the coast (Gallardo 2009). Both of thesepatterns are evident in sites such as those discussedhere, sites that occupy “empty” spaces betweenpopulations and foment interregional interaction.

During archaeological survey along the San Sal-vador River, the remains of a small village andcemetery were documented approximately 110 kminland along the river at 1,370 m asl (Figure 1).Other than a brief note on a looted tomb (Spahni

1964), this survey provided the first evidence of pre-historic human occupation along the San Salvador.The San Salvador sites are found 40 km west ofCalama in empty desert close to a prehispanic routethat linked the Calama oasis west to the coast atCobija (110 km) and northwest to the Quillaguaoasis (93 km). Their size, specialized production,and material culture suggest their logistical char-acter in service to the population of Calama. Thisis particularly true if one considers that the archae-ological localities mentioned above have Forma-tive period cemeteries that have yielded hundredsof human remains. These suggest increasing pop-ulation resulting from sedentism and the intensifi-cation of production during the Formative (e.g.,Gallardo et al. 1993; Moragas 1982).

Excavation and surface collection were con-ducted at the San Salvador Village and Cemeteryin 2008. Radiocarbon dates place both sitessquarely in the Middle Formative period (cemetery:2080 ± 40 B.P. [Beta-247417], 164 cal B.C.–calA.D. 128 [p = .07] and 121 cal B.C.–cal A.D. 67[p = .93]; village: 2370 ± 40 B.P. [Beta-247418],511–350 cal B.C. [p = .72] and 311–209 cal B.C.[p = .28]; calibrated at 2! with the program CALIB6.0 using the SHCal04 curve [McCormac et al.2004; Stuiver et al. 2005]). The village was datedusing a roof tie found in one of the rooms, offer-ing evidence for the earliest portion of the Salvadoroccupations. This suggests that the two dates areconsistent, given that the cemetery, which is spa-tially and culturally associated with the village,should postdate the original construction of thehabitations.

The San Salvador Village (Aldea San Salvador),located on a terrace above the river west of thecemetery, comprises approximately 15–20 circu-lar pits with poles sustaining thatched roofs. Thisstyle of construction is consistent with the struc-ture of other habitation sites along the Middle LoaRiver (Pollard 1970). Excavation of a trash mid-den confirmed the presence of camelids and anunusual abundance of algarrobo seeds (Prosopisflexuosa), a local fruit, as well as the remains ofmarine fish, likely transported to the site as a driedgood. Multiple species of marine fish (Trachurussimetricus, Cilus gilberti, and Graus nigra) andshells, including some carved into small beads thatcould have been used as adornments (Turritela cin-gulata, Oliva peruviana, Loxechinus albu, and

Torres-Rouff et al.] EATING FISH IN THE DRIEST DESERT IN THE WORLD 53

Choromytilus chorus), were found at the site, sug-gesting a tie to coastal populations. Our excavationsrevealed no remains of freshwater fish. Finds at thevillage site also included vegetable and woolcordage, leather, feathers, a projectile point, carvedwooden instruments, cactus spines, and gourds. Itis probable that the settlement was placed at thispart of the river’s course because of its proximityto grazing lands and to a small forest of algarrobotrees (de Ugarte and Gallardo 2010). The incredi-ble preponderance of algarrobo among the botan-ical remains in concert with the presence ofgrinding stones suggests the site’s role as a pro-duction site for algarrobo flour, which could thenbe used as an exchange good as well as serving asa dietary staple. Foreshadowing the isotopic resultsdiscussed below, no evidence of maize (Zea mays)was recovered in the cemetery or residential exca-vations (de Ugarte and Gallardo 2010) despite 5-mm mesh screening and its archaeological visi-bility at other sites in the region, where it is fre-quently included as a grave offering.

The San Salvador Cemetery (Cementerio SanSalvador) is located just upriver from the villagesite. Found on an elevated ridge above the riveritself, the cemetery likely comprises some 20–25interments in small ovoid pits demarcated bywooden posts. Some portion of the cemetery waslooted during the nitrate boom of the 1930s, asdetermined through glass bottles found at the site.As a result, many human remains were exposed andcommingled on the surface. We recovered allhuman remains and remaining artifacts from eightof these tombs as well as the surrounding areas.Tombs appear to have included, among otherthings, pottery, baskets, lithics, beads, and textilesin the form of clothing and bags. Also providingevidence for a coastal tie, loco (Concholepas con-cholepas) shells filled with pigment were amongthe grave goods. In line with the finds at the villagesite, we found a small bundle of algarrobo seed-pods tied together with cord among the offeringsin one tomb but no evidence of maize, a commoninclusion in other regional graves. Contemporaryburials from sites in Calama have produced mate-rial culture (including textiles, ceramics, and bas-ketry) that parallels that seen in the mortuarycontexts at San Salvador (González and Westfall2006; Thomas et al. 1995). However, reflecting theFormative period’s patterns of increasing interre-

gional interaction, goods found at the San SalvadorCemetery come from diverse areas.

Considered together, these excavations yieldedevidence of interaction not just with the coast asdetailed above but also with northwest Argentina.Among the trans-Andean finds were the leg of aceramic figurine, a tubular black polished pipe frag-ment, and black polished ceramic fragments inshapes similar to those found in the Calchaquí Val-ley (e.g., Baldini 2007). In concert with the coastalmaterial, these archaeological details support theidea that the San Salvador sites were logistical out-posts that facilitated movement between the coastand the interior. Evidence from the site indicatesintensive exploitation of the valley’s plantresources, maintenance of camelid herds, regularaccess to coastal products, and a close relationshipwith contemporary populations from the coastalareas of Cobija and the mouth of the Loa River aswell as with other oases, such as Quillagua, Calama,and Chiu-Chiu (Agüero et al. 2004; Moragas 1982;Pollard 1970; Spahni 1964, 1967; Thomas et al.1995). Below we present the results of skeletalanalyses of the 17 fragmented human remainsrecovered from the San Salvador Cemetery toexplore the inhabitants of this area in more detail.We follow this with a detailed analysis of paleodietas determined through stable isotope analyses ofcarbon and nitrogen.

Skeletal Analyses

The remains of 17 individuals were recovered fromeight tombs and surrounding areas at the San Sal-vador Cemetery (Table 1). All human skeletalremains were assessed using standard bioarchaeo-logical protocols (Buikstra and Ubelaker 1994;Buzon et al. 2005). Minimum number of individu-als was calculated for each of the tombs and sur-rounding areas. Sex was determined based on thesexually dimorphic features of the os coxae and theskull. Similarly, broad age assessments were madebased on pelvic morphology, epiphyseal fusion, andcranial suture closure. Remains were also analyzedand scored for visible pathology and trauma. Allobservable pathology was documented using stan-dard methods and terminology (Buikstra and Ube-laker 1994). Finally, each skull was examinedvisually for the presence and type (annular vs. tab-ular) of deliberate cranial shaping using adjustments



Tabl

e 1.

Indi

vidu

als

Rec

over

ed fr

om th

e Sa

n Sa

lvad

or C

emet

ery

and

Res

ults

of S

tabl

e Is

otop

e A

naly

ses.

Indi

vidu

als

in b

old

and

italic

s w

ere

excl

uded

from

con

side

ratio

n du

e to

evi

denc

e of

con

tam

inat

ion,

deg

rada

tion,

or i

nstru

men

t err

or.

For c

olla

gen

anal

yses

, rep

licat

e an

alys

es o

f ace

tani

lide

(the

labo

rato

ry s

econ

dary

sta

ndar

d), a

nd w

orki

ng s

tand

ards

(thi

oure

a an

d hy

drox

ypro

line)

resu

lted

in re

prod

ucib

ility

of ±

0.1‰

for !

13C

and

±0.

2‰ fo

r !15

N. F

or

hydr

oxya

patit

e ca

rbon

ate

anal

yses

, rep

licat

e an

alys

es o

f BaC

O3 a

nd C

aCO

3 res

ulte

d in

repr

oduc

ibili

ty o

f ±0.

1‰ fo

r !13

C. C

arbo

n is

otop

e ra

tios

are

repo

rted

rela

tive

to th

e V

-PD

B (V

ienn

a Pe

eDee

bel

emni

te) c

arbo

nate

st

anda

rd a

nd a

re e

xpre

ssed

in p

arts

per

mil

(‰) u

sing

the

follo

win

g st

anda

rd fo

rmul

a: !

13C

= ((

(13C

/12C

sam

ple)/

(13C

/12C

stan

dard

)) –

1) "

1,00

0 (C

raig

195

7). N

itrog

en is

otop

e ra

tios

are

repo

rted

rela

tive

to th

e A

IR (a

tmos

pher

ic

nitro

gen)

sta

ndar

d an

d ar

e ex

pres

sed

in p

arts

per

mil

(‰) u

sing

the

follo

win

g fo

rmul

a: !

15N

= ((

(15N

/14N

sam

ple)/

(15N

/14N

stan

dard

)) –

1) "

1,00

0 (M

ario

tti, 1

983)

.

In

divi

dual

A

ge

Sex

Isot

ope

sam

ple

no.

Sam

ple

type

El

emen

t sam

pled

C

olla

gen

yiel

d (w

t%)

w

t% C

wt%

N

Ato

mic

C

:N

!15N

co

(‰)

!13C

co

(‰)

Apa

tite

yiel

d (w

t%)

!13C

ap

(‰)

#13C

ap-c

o (‰

)

1.1

Adu

lt M

ale?

A

-19

Bon

e

Rib

4.

4 13

.5

40.0

3.

5 11

.2

-16.

6 70

.2

-10.

9 5.

8 1.

2 Ju

veni

le (8

-12

yo)

- A-

20

Bone

Ce

rvic

al v

erte

bra

1.1

- 2.

1 -

- -

- -

-

F-26

To

oth

(roo

t) Ri

ght m

axill

ary

P4

1.5

- 0.

8 -

- -

91.1

-2

0.3

-

E-10

9 B

one

4th

met

atar

sal

2.7

4.1

12.4

3.

5 13

.3

-17.

4 74

.1

-13.

3 4.

2 1.

3 A

dult

Fem

ale?

n/

a n/

a n/

a n/

a n/

a n/

a n/

a n/

a n/

a n/

a n/

a n/

a 1.

4 Ju

veni

le (4

-6 y

o)

n/

a n/

a n/

a n/

a n/

a n/

a n/

a n/

a n/

a n/

a n/

a n/

a 1.

5 Ju

veni

le (2

-4 y

o)

n/

a n/

a n/

a n/

a n/

a n/

a n/

a n/

a n/

a n/

a n/

a n/

a 1.

6 Fe

tal

n/

a n/

a n/

a n/

a n/

a n/

a n/

a n/

a n/

a n/

a n/

a n/

a 2.

1 A

dult

Fem

ale

A-21

Bo

ne

Righ

t rad

ius

3.2

- -

- -

- -

- -

A-

22

Bone

1s

t rib

1.

6 -

2.3

- -

- 85

.9

- -

2.2

Adu

lt

A-23

Bo

ne

Rib

0.6

4.4

14.1

3.

7 12

.3

-18.

2 -

- -

A-

71

Bone

Cl

avic

le

0.7

- 8.

5 -

- -

77.1

-

-

F-27

To

oth

(roo

t) M

3?

12.3

12

.8

36.2

3.

3 15

.4

-16.

9 88

.5

-12.

1 4.

9 3.

1 M

iddl

e A

dult

Mal

e A

-25

Bon

e H

umer

us?

2.1

11.5

34

.2

3.5

12.9

-1

8.3

82.0

-1

0.4

7.9

5.1

Mid

dle

Adu

lt M

ale

A-2

6 B

one

Rib

19

.0

14.5

42

.1

3.4

14.4

-1

5.5

63.3

-1

0.6

4.9

5.2

Adu

lt

A-2

7 B

one

Cra

nial

frag

men

t 5.

1 13

.0

37.8

3.

4 14

.6

-16.

2 71

.8

-11.

3 4.

9 8.

1 A

dole

scen

t (16

-20)

A-28

Bo

ne

Rib

0.4

- -

- -

- -

- -

A

-70

Bon

e ui

d lo

ng b

one

4.0

10.8

29

.9

3.2

8.34

-1

7 78

.3

-9.9

7.

1

F-25

To

oth

(ena

mel

) M

3 -

- -

- -

- 96

.6

-14.

3 n/

a 9.

1 Y

oung

Adu

lt

n/a

n/a

n/a

n/a

n/a

n/a

n/a

n/a

n/a

n/a

n/a

n/a

10.1

A

dult

A

-29

Bon

e H

umer

us?

0.6

7.2

22.4

3.

6 10

.6

-17.

4 83

.4

-9.7

7.

8

A-7

2 B

one

Fibu

la

1.9

5.8

17.2

3.

5 11

.6

-15.

5 79

.9

-9.4

6.

1 11

.1

Juve

nile

(4-6

yo)

A-30

Bo

ne

12th

rib

1.0

- 3.

2 -

- -

- -

-

E-11

0 Bo

ne

Rib

1.7

2.6

8.5

3.9

12.6

-1

6.9

72.1

-

-

to the broad classification system that originateswith Dembo and Imbelloni (1938). In all cases,preservation differentially affected the sample.

The San Salvador skeletal material is primarilypostcranial, likely the result of the aforementionedlooting. The 17 individuals recovered from SanSalvador Cemetery ranged in completeness fromindividual 5.1, represented by very nearly an entireskeleton, part of which was naturally mummified,to several individuals represented by minimal frag-ments (i.e., individuals 6.1 and 7.1). While a num-ber of remains could not be sexed or aged (or couldonly be determined to be adults), analysis of thesample revealed a number of juveniles (n = 6) aswell as members of both sexes (Table 1).

Pathology in the sample appears to have beenlimited to degenerative processes (osteoarthritis)and traumatic injury. There is no evidence of infec-tious disease in any of the skeletal remains. Nev-ertheless, the remains seem to show effects fromliving in a harsh environment. Qualitative appraisalof muscle insertion sites and musculoskeletal stressmarkers on available adult long bone fragmentssuggests pronounced robusticity and perhaps sub-stantial muscle use for two of the three adult malesin this population (Hawkey and Merbs 1995).1

Given that these results may be a reflection of long-term body use, they might be considered a productof frequent activities in difficult terrain as opposedto a sign of significant labor intensity (e.g., Niin-imäki 2011). Osteoarthritis is visible on most adultremains; however, vertebrae were the only affectedelements. This takes the form of coalescing poros-ity and lipping on vertebral bodies and articular sur-faces, except in the case of individual 5.1, wherethis is augmented with the formation of osteophytesand the fusion of the articular surfaces of the firstand second cervical vertebra.

Other pathological changes noted in this sam-ple result from traumatic injury. All three adultmale individuals whose remains were more com-plete showed evidence of fractures. Some injurieswere quite significant, including a healed mandibu-lar fracture below individual 5.1’s left mental fora-men (Figure 2). Individual 3.1 suffered a series ofcomplete fractures to the ribs ultimately resultingin misalignment of the ends during healing, pro-ducing substantial overlap medially (Figure 3).Individual 1.1 also showed healed rib fractures aswell as injuries of his left distal ulna and radius,

resulting in what appears to be a nonunion fractureof the distal metaphysis of the ulna. While thiscould be a parry fracture, a defensive injury (Buik-stra and Ubelaker 1994; Judd 2008), the involve-ment of the radius suggests otherwise. It has beenargued that cranial injuries are more indicative ofintentional injury or violence, but there is no clearevidence of interpersonal violence among theseindividuals. It is possible that these injuries areproducts of the rocky landscape at San Salvador orsimilar areas. Unsurprisingly, analyses of cranialvault modification were hampered by the lack ofcrania. No cranial fragments could be definitivelyassessed for alterations to head shape. However, theone complete cranium (individual 5.1) demon-strated tabular erect cranial modification with asmall degree of pressure resulting in a slight flat-tening of the anterior and posterior aspects of theskull (Figure 4). This form of head shaping is notunusual in north Chile’s interior (Torres-Rouff2007). Given that cranial modification results fromthe inscription of cultural signifiers on the bodiesof infants, it reflects strong, and likely lessephemeral, associations and as such may be sug-gestive of long-standing familial ties to a particu-lar place or group in the interior.

While the skeletal remains do not provide directevidence of origin or exchange, a few salient piecesof information can be derived from them. As men-tioned above, our demographic data reveal the pres-ence of potential family groups at the cemetery,suggesting a degree of permanence and stability forthis settlement. This is bolstered somewhat by pat-terns of osteoarthritis and traumatic injury, both ofwhich intimate a long-standing relationship withthe landscape of the San Salvador Valley. The onecomplete cranium showed evidence of tabular erectmodification, which is also suggestive of inland ties.During the Formative period, tabular forms are themost common shape seen among populations in theinterior, and they stand in sharp contrast to the annu-lar (circumferential) forms that are typical of con-temporary coastal sites (Torres-Rouff 2003, 2007).

Stable Isotope Analyses

Isotopic reconstruction of human paleodiet is pred-icated on known variation in the carbon and nitro-gen isotope signatures of broad classes of plant andanimal foodstuffs. Incorporation of these known-


value items into body tissues allows for the recon-struction of the relative contribution of each class offoodstuff, ultimately providing individually specificdata about diet. In cases where the dominant foodsvary between regions (e.g., marine foods at the coastvs. terrestrial foods in inland areas), these data canalso be used as a proxy for assessing mobility.

Stable carbon isotopes in plants vary accordingto photosynthetic pathway. Most terrestrial plants(including many cereals, legumes, vegetables, nuts,and fruits and locally available staples like potato andalgarrobo) use a C3, or Calvin, photosynthetic path-way; tropical grasses like maize (Zea mays) and thepseudocereal quinoa (Chenopodium spp.) use the C4,


Figure 2. Healed fracture of left mandible (Individual 5.1).

Figure 3. Misaligned healed rib fracture (Individual 3.1).

or Hatch-Slack, photosynthetic pathway; and cactiand succulents use the crassulacean acid metabolism(CAM) pathway (Calvin and Benson 1948; Hatchand Slack 1966; Kortschak et al. 1965; Ranson andThomas 1960). Carbon isotope signatures also dif-fer significantly between organisms from marineand terrestrial ecosystems as a result of dramatic dif-ferences in their carbon sources (Chisholm et al.1982; Schoeninger and DeNiro 1984).

Stable nitrogen isotopes vary in a fairly pre-dictable manner according to trophic level and areparticularly useful for identifying marine food con-sumption given the clear differences in marine andterrestrial organisms’ nitrogen isotope signatures(Minagawa and Wada 1984; Schoeninger andDeNiro 1984; Schoeninger et al. 1983), althoughthere are also climatic factors, aridity being mostpertinent here, that can affect nitrogen isotope val-ues (Ambrose 1991; Ambrose and DeNiro 1987).Since both plants and animals vary from region toregion in their carbon and nitrogen signatures as aresult of a number of environmental factors, localisotopic signatures are a necessary part of pale-odietary reconstructions.

Turning to the human consumers, different frac-tions of skeletal or dental tissues reflect differentportions and/or different temporal dimensions ofan individual’s diet. Carbon isotope signatures inbone apatite ("13Cap) reflect carbon sources in thewhole diet (these are predominantly carbohydrate-based in humans).2 Carbon and nitrogen isotopesof bone collagen ("13Cco and "15Nco) overwhelm-ingly reflect dietary protein sources (Ambrose andNorr 1993; Kellner and Schoeninger 2007). Whilethe chemical makeup of bone collagen and hydrox-yapatite reflects average diet over the last 10–30years of an individual’s lifetime, the isotopic sig-nature of tooth root (dentine) collagen and enamelapatite reflects diet at the time of a particular tooth’sformation. Therefore, chemical analyses of differ-ent body tissues and biomolecules provide insightsinto diet and mobility data over different portionsof an individual life.

For this study, bone collagen, tooth root (den-tine) collagen, and bone, tooth root (dentine), andenamel hydroxyapatite were extracted and purifiedin the National Science Foundation Stable IsotopePreparation Lab at the Field Museum of Natural


Figure 4. Lateral view of cranium (Individual 5.1).

History using protocols from the EnvironmentalIsotope Paleobiogeochemistry Laboratory at theUniversity of Illinois, Urbana-Champaign(Ambrose 2007a, 2007b). These protocols arederived from Longin (1971) and Ambrose (1990)for dentine and bone collagen and Lee-Thorp(1989) and Koch et al. (1997) for dentine, enamel,and bone hydroxyapatite. Tooth root (dentine) col-lagen and apatite samples and enamel apatite sam-ples were derived from complete homogenizedroots or crowns rather than being sequentiallymicrosampled. All isotope analyses were per-formed in the Stable Isotope Laboratory of theDepartment of Earth and Environmental Scienceat the University of Illinois at Chicago. Collagenanalysis was carried out using a Thermo-FinniganDeltaPLUSXL stable isotope ratio mass spectrome-ter equipped with a Costech Analytical ECS 4010Elemental Analyzer coupled to the mass spec-trometer via a Thermo-Finnigan ConFlo III.Hydroxyapatite carbonate samples were introducedto the mass spectrometer using a Thermo-FinniganGasBench II.

The observed variation in macroscopic preser-vation of the human remains from San Salvadorwas mirrored in substantial disparities in the fourchemical and elemental indexes most commonlyemployed in evaluating collagen preservation: col-lagen yield, carbon yield (wt% C), nitrogen yield(wt% N), and atomic C:N ratio (Table 1; Ambrose1990; van Klinken 1999). Collagen preservationmust be assessed before dietary reconstruction isattempted since postmortem changes to bone canproduce erroneous chemical signatures. The aver-age collagen yield of the 16 bone samples in thestudy was quite low (3.1 ± 4.5 wt%), and one sam-ple, A-28, had insufficient collagen for further con-sideration. A further five bone samples (A-20, A-21,A-22, A-30, A-71) had insufficient carbon (< 4.5wt%) and/or nitrogen (< .5 wt%) and could not beconsidered chemically unaltered. Finally, weexcluded two additional samples (A-23, E-110)with acceptable collagen yields and elemental con-centrations because of atomic C:N ratios that wereoutside of the acceptable range (2.9–3.6). All told,only eight of the original 16 bone samples (50 per-cent) were sufficiently well preserved for consid-eration in dietary reconstruction. While both dentalroot (dentine) samples for which collagen extrac-tion was attempted produced sufficient collagen

(1.5 wt% and 12.3 wt%), the sample with the lowercollagen yield (F-26) did not produce sufficientconcentrations of carbon and nitrogen. As such,the isotopic value of collagen from only one den-tal root (dentine) sample (F-27) was considered fordietary reconstruction. Extraction of the solitarycrown enamel apatite sample (F-25) produced val-ues within expected norms. We did not performFourier-transform infrared spectroscopy, which hasbeen shown to be effective in independently iden-tifying diagenetic alteration of hydroxyapatite(Trueman et al. 2008). Instead, we employed acombination of the above-mentioned bone colla-gen preservation measures (Bösl et al. 2006;Schutkowski et al. 1999) and compared the isotopicvalues of bone and tooth apatite to evaluate thestate of apatite preservation.

To permit dietary reconstruction from theremaining samples, we assembled a representativefood web for northern Chile from published iso-topic data for local flora and fauna (DeNiro andHastorf 1985; Horn et al. 2009; Hückstädt et al.2007; Minagawa and Wada 1984; Schoeninger andDeNiro 1984; Schoeninger et al. 1983; Tieszen andChapman 1992) and from analysis of a small num-ber of ancient and modern floral samples (includ-ing algarrobo) collected in or around the SanSalvador sites. The isotopic signatures in the recon-structed ancient food web (Figure 5) can effec-tively be divided into three mostly exclusivesubsets: (1) C3 plants and terrestrial fauna, mostnotably represented by algarrobo and camelids,respectively, which possess generally depleted "13Cand "15N signatures;3 (2) C4/CAM plants, the "15Nvalues of which are roughly equivalent to the firstgroup but which possess significantly moreenriched "13C values; and (3) marine fauna, suchas the fish and shell species found at the site, whichpossess dramatically enriched "13C and "15N sig-natures. Based on these local differences, it shouldbe possible to differentiate both C3 vs. C4/CAMplant consumption and marine vs. terrestrial pro-tein consumption for the San Salvador individuals.

Putting these elements together, we start byreconstructing the whole diet of the San Salvadorindividuals using their "13Cap values (Table 1). Onaverage, the "13Cap of bone and enamel reflects thecarbon isotope values of all plants and animals con-sumed, plus 9.4‰ (Ambrose and Norr 1993), butsince carbohydrates tend to dominate in human diet


these results are viewed as a reflection of plant con-sumption. The observed range, from a low of–14.3‰ (tooth enamel from individual 8.1) to ahigh of –9.4‰ (bone from individual 10.1), speaksto significant differences in long-term patterns of(primarily) plant consumption. This 5‰ variation,when compared against local plant end-member(C3 vs. C4/CAM) values that differ by an averageof 11‰ (–25‰ vs. –14‰), could equate to as muchas a 40 percent shift from C3 to C4/CAM plant con-sumption.4 Put differently, individuals with themost depleted "13Cap values would appear to haveconsumed predominantly C3 plants with C4/CAMplants only providing a minor contribution to diet,whereas those individuals with the most enriched"13Cap signatures had a diet consisting of as muchas 40 percent C4/CAM plants. It is noteworthy,however, that none of the individuals analyzed inthe course of the present study have "13Cap valuesthat approach the enrichment seen in, for instance,Middle Horizon (A.D. 500–1000) individuals fromConchopata, Peru (Finucane et al. 2006), southcoastal Peru (Nasca [Kellner and Schoeninger

2008]), certain individuals from central Chile (Fal-abella et al. 2007), or other C4 carbohydrate–depen-dent populations presented by Kellner andSchoeninger (2007). This finding is consistent withthe lack of archaeological recovery of C4 foodstuffsin the San Salvador Village and Cemetery (deUgarte and Gallardo 2010), and its implications arediscussed below.

An intriguing finding from these data, however,is that the three samples reflecting the highest lev-els of C3 plant consumption were from juveniles(bone from individual 1.2) and/or were dental (rootdentine and crown enamel) apatite samples reflect-ing a period of childhood/adolescent diet (M3 sam-ples from individuals 2.2 and 8.1).5 On the otherhand, the bone samples from adults at the sitereflect a more diversified carbohydrate diet. Giventhat the overwhelming majority of the botanicalremains from San Salvador were from a C3 plant(algarrobo, "13C –21.1‰ to –22.8‰), these datamay suggest a pattern in which children, perhapsborn to families residing at the site, consumed thelocally available carbohydrate (C3) throughout


Figure 5. San Salvador foodweb.

childhood/ adolescence, while C4/CAM carbohy-drates were only introduced into diets later in life.This may be tied to some form of greater adultmobility, if one assumes that local food remainsare indicative of local diet, or it may suggest restric-tions on juvenile diet. In either case, it provides aninteresting view of the effect of food exchange onindividual actors.

Turning from the carbohydrate to the proteinportion of the diet, both the measured "13Cco val-ues and the difference between the measured apatiteand collagen "13C “spacing” values (#13Cap-co) canbe used to determine, in broad terms, the "13C valueof the predominant dietary protein (Ambrose andNorr 1993). Measured "13Cco values for the San Sal-vador individuals average –16.8‰ and range from–15.5‰ to –18.8‰, whereas the calculated #13Cap-covalues average 5.9‰ and range from 4.2‰ to7.9‰. Taken together, these values suggest an aver-age dietary protein "13C signature between thedepleted "13C values (< –20.5‰) of local terres-trial fauna and the enriched "13C signatures(>–14.8‰) seen for regional freshwater fish andmarine taxa. That the spacing values averagearound 5‰ likely attests to the fact that dietaryproteins, like the dietary carbohydrates discussedabove, were drawn from multiple, isotopically dis-tinct reservoirs, with some individuals, particularlythose with #13Cap-co values below 5‰, consumingslightly more marine resources and those with val-ues greater than 5‰ eating more terrestrial protein.This final inference is lent some support by themoderate (r2 = .42) negative correlation observedbetween #13Cap-co and "15Nco.

Coupling these results with the measured "15Nvalues of collagen (average 12.5‰, range8.3‰–15.4‰) allows us to make further statementsabout the protein diet of the San Salvador individ-uals. Based on knowledge of the local food web(Figure 5) and the roughly +3‰ stepwise enrich-ment of "15N values from one trophic level to thenext, "15Nco values greater than 12.5‰ can only beexplained by the consumption of 15N-enrichedmarine fauna. Five of the nine San Salvador indi-viduals have "15Nco values exceeding this 12.5‰threshold. Some portion of the observed variationin "15Nco may result from a lingering elevation oftrophic positioning occurring as a consequence ofbreast-feeding in younger individuals (e.g., Fulleret al. 2006). Given, however, that the children

included in this analysis were either beginning ado-lescence or represented by samples that reflect ado-lescence (Richards et al. 2002), we argue that thesedata are not a consequence of a strong breast-feed-ing effect and instead support the presence of 15N-enriched marine fauna in the San Salvador diet.

Much like the patterns seen in the carbohydrateportion, what is evident from the collagen data, andparticularly the "15Nco values, is a substantial degreeof intrasocietal variation in protein consumption.Interestingly, isotopic analyses of collagen fromtwo individuals from the contemporary Middle LoaRiver site of Regimiento Chorillos (820–390 B.C.)suggest similar variation in dietary protein(González and Westfall 2006). If the observed vari-ation in collagen isotope values is the result of bonafide differences in post-weaning protein consump-tion, then it is most likely that this results from dif-fering amounts of marine vs. terrestrial protein inthe diet of San Salvador individuals. This is alsosupported by the archaeological recovery of marinefauna from the site. Using local "15N end-membervalues of terrestrial/freshwater (average "15N =6.7‰) and marine (average "15N = 18.4‰) faunaand the formula for computing marine consump-tion from "15Nco provided by Ambrose (1993:83),we find a 45 percent range in marine protein con-sumption, from as little as 6–12 percent marineconsumption (in individuals 8.1 and 10.1) to asmuch as 45–51 percent marine consumption (indi-viduals 5.2 and 2.2).6 The positive, but admittedlyweak, correlation (r2 = .16) between "15Nco and"13Cco values and the moderate (r2 = .42) negativecorrelation observed between #13Cap-co and "15Ncoboth provide some support for this interpretation.As with the "13Cap data, this idiosyncratic variationin collagen isotope signatures may reflect intraso-cietal differences in patterns of mobility (i.e., fre-quent movement to and from the coast by someindividuals vs. greater sedentism for others) or asa result of the importation of marine foods (e.g.,dried fish) over great distances to San Salvador. Thelack of a uniform contribution of marine protein inall individuals’ diets may suggest the interaction ofthese two patterns, with certain individuals involvedmore heavily in trade relationships than others.

For some regional perspective on these results,we compared the collagen isotope values from SanSalvador with those from individuals from otherAndean sites, including six from north coastal Chile


(Tieszen et al. 1992), nine from central Chile (Fal-abella et al. 2007), 16 from central westernArgentina (Gil et al. 2006), and two in neighbor-ing Peru (Finucane et al. 2006; Kellner andSchoeninger 2008). These fall into three broadgroupings (Figure 6): first, the sites from centralChile and central western Argentina, which demon-strate a heavy reliance on C3/terrestrial proteins;second, individuals from the Peruvian sites (bothhighland and coastal), who appear to have con-sumed mixed C3/C4 and C4-foddered terrestrial pro-tein sources; and third, the individuals from coastalnorthern Chile, who exhibit patterns of protein con-sumption focused on marine resources.

When assessing our collagen results vis-à-visthese regional counterparts, two phenomena areevident. First, the "13Cco values of the San Salvadorindividuals are most similar to the values observedin central Chile and west-central Argentina. In thesethree areas the "13Cco signatures of almost everyindividual from the comparative populations areless than –14‰, the dividing line between C3 andC4/marine protein established by Kellner andSchoeninger (2007) in their recent synthesis ofnumerous controlled diet experiments. Interest-

ingly, the groups from Peru and north coastal Chileshow values that are greater than –14‰. At San Sal-vador "13Cco values average –16.8‰, indicating aprotein diet slanted toward C3/terrestrial protein.We interpret this difference in "13Cco signatures tobe a consequence of two differences in dietary com-position: (1) the direct human consumption of farfewer C4 plants at San Salvador than at the othersites (a conclusion buttressed by the dearth of C4plant material recovered from San Salvador); and(2), of more significance to collagen isotope val-ues, the dominance of C3-foddered terrestrial faunain the protein portion of the average San Salvadordiet, with the protein diet at the comparative sitesbeing made up of varying combinations of marinefauna and C4-foddered terrestrial fauna in additionto C3-foddered fauna.

Furthermore, the "15Nco values of the San Sal-vador individuals are equal to or greater than thevalues seen in individuals from central westernArgentina, central Chile, and the Peruvian sites, andthey approach the lower end of values seen in indi-viduals from north coastal Chile. As discussedabove, based on local food web values, the inter-mediate trophic position of the San Salvador indi-


3

8

13

18

23

-20 -18 -16 -14 -12 -10 -8

!15 N

co (‰

)

!13Cco (‰)

San Salvador

Coastal Peru (Nasca)

Highland Peru (Conchopata)

North Coastal Chile

Central Western Argentina

Central Chile

23

18

Nco

(‰)

dor S valan SaS

aC ru (Ne PlatsCoa

) acsa

8

13 !15 N

ru (ConcH end PaghliH

ChiN latsh CoaortN

AC rn etseWWlrantCe

eC l ChilrantCe

) athoparu (Conc

el Chi

naintgerA

3

-20

-18 -16

-14

!13Cco (‰)

-12 -10

-8

Figure 6. Collagen isotope values of San Salvador humans and comparative regional populations.

viduals is not a consequence of aridity-relatedenrichment of terrestrial "15N values. Rather, weargue that at least some of the San Salvador indi-viduals (those with higher "15Nco values) were con-suming more protein of a higher trophic level (i.e.,marine protein) than the individuals from the Peru-vian, central Chilean, and Argentina sites but lessthan most of the individuals from the north coastalChilean sites, whom Tieszen and colleagues (1992)argue were consuming a minimum of 45–50 per-cent marine protein.

Combining these two pieces of collagen isotopeevidence, we are able to arrive at a parsimoniousexplanation for a peculiar pattern of protein (andcarbohydrate) consumption at San Salvador thatwould produce depleted "13Cco values and moder-ately enriched "15Nco values. The collagen isotopevalues observed in central western Argentina andcentral Chile reflect a diet of mixed C3/C4 carbo-hydrates and predominantly C3/terrestrial protein.Individuals from the Peruvian sites appear to haveconsumed primarily C4 carbohydrates plus C4-fod-dered terrestrial fauna, thus producing the enriched"13Cco and depleted "15Nco signature combination.7

Individuals from the comparative north coastalChilean sites likely were also consuming a largeamount of C4 carbohydrates (although we do notpossess apatite data for these individuals) in addi-tion to varying combinations of C4-foddered ter-restrial fauna and marine animals, therebyproducing their combination of enriched "13Cco val-ues and "15Nco signatures that range from moder-ately to exceedingly enriched. In contrast, at SanSalvador, we see a distinct combination of primar-ily C3 dietary carbohydrates (with, at most, a 40percent minority of C4 carbohydrate consumptionin some individuals) and a protein diet made up ofmostly C3-foddered terrestrial fauna plus varyingamounts (5–50 percent) of marine protein. Weargue that it is this combination of foodstuffs thatresulted in depleted "13Cco and moderately enriched"15Nco values at San Salvador.

In summary, these isotopic results concord wellwith both the archaeological and the bioarchaeo-logical data from San Salvador. Our results suggestthat the inhabitants of San Salvador primarily con-sumed the dominant local carbohydrate, algarrobo,in particular during their childhood, and once theygrew older, took part in interregional trade expedi-tions, thereby gaining access to and consuming

varying amounts of coastal and highland foodresources (meat and plant alike). As a consequenceof this pattern of lifetime consumption, the inhab-itants of San Salvador possessed dietary signaturesthat span the “isotopic space” between those mostoften seen in the coast and the interior, respectively.

Discussion

While the quality of this sample does not allow usto make grand statements about life during the For-mative period in the Atacama Desert, we argue thatthere is some tantalizing evidence suggesting pat-terns of interaction for this population. Not only dowe see a space in between population groups inter-acting with peoples and objects from diverse areas,but also our data are suggestive of the movementof these peoples: “This raises the question of thesignificance of distant worlds beyond daily inter-action for the constitution of both individual andcollective identities” (Lazzari 2005b:130). Theinhabitants of San Salvador created a social andeconomic space within the spectrum of interac-tions they experienced. The location and materialremains contribute to the argument for a strongconnection to the peoples of the Middle Loa River.As mentioned earlier, there is a demonstrated pres-ence of textiles, baskets, and ceramics akin to thosefrom the interior in addition to material from dis-tant cultural groups. For example, a textile fragment(Figure 7) from San Salvador bears strong resem-blance to a contemporary one found at the Calamaoasis site of Topater (2320 ± 40 B.P. [Beta-259695],410–360 cal B.C.).

Archaeological evidence shows that the San Sal-vador population had clear relations with the inte-rior, but in concert with our analyses it also showsthat this group had a strong tie with contemporarycoastal populations, despite the site’s location deepin the Atacama Desert. For example, we note thepresence of numerous marine fish remains, as wellas marine shells, in the site middens. Given thesite’s location over 100 km from the Pacific Oceanthis argues for the involvement of San Salvador’sresidents in resource exchange with coastal popu-lations, probably through the trading of algarroboflour, some of which we found in baskets interredas grave goods at San Salvador Cemetery. Faunalremains (principally paleofeces) recovered duringour excavations suggest the presence of camelids


at the site— animals that likely played an impor-tant role in these relations. As noted before the siteswere located near a grove of algarrobo trees alongthe San Salvador River. Jochim (2006:88) hasargued that the presence of sites in relationship tospecific resources can serve as one piece of evi-dence for intensification and the presence of foodexchange. Taken together these observations sup-port de Ugarte and Gallardo’s (2010) suggestionthat the site was placed in this location for pro-duction purposes.

While it is likely that this population had its ori-gins near the modern inland city of Calama, approx-imately 40 km east of the sites, our evidencesuggests not only that it served as a geographicintermediary but that its relationships with inlandand coastal populations were complex and varied.As Lazzari and colleagues note, these “complexsocial networks . . . extended people’s involvementbeyond their daily locations into the wider regional

world” (2009:1963). The peoples of San Salvadorhad access to seemingly exotic goods in the formof both material culture and marine foodstuffs.Their homes and graves evidence their local affil-iations and their interactions with a broader world.As alluded to above, archaeological and bioar-chaeological data converge on a pattern of greatvariety for this population. Interestingly, the iso-topic data reveal a novel combination of food in thediet when compared with available data from theregion. While the absence of evidence for maizeamong the botanical remains and cemetery offer-ings is not evidence of complete absence, it is sug-gestive of a minimal consumption of corn, a typicalAndean staple that was found among the burialofferings at the contemporary cemetery of Topaterin the Calama oasis (Thomas et al. 1995). Regard-less of the specific foodstuffs involved, these datasuggest different dietary experiences and life his-tories for the individual occupants of the site,


Figure 7: Textile fragment from San Salvador cemetery (R) and similar fragment from Loa River site of Topater(Calama) (L).

including the potential for different places of ori-gins, varied durations of residency at San Salvador,and variation in the frequency of long-distancetravel and trade. Moreover, these data suggest inter-esting intrasocietal patterning that may reflect dif-ferential access to resources based on age, sex, orsocial standing. Given the seeming stability of thesettlement and the abundance of marine remains,it appears that this group had somewhat regularaccess to these products through exchange withcoastal populations moving inland along the courseof the Loa and San Salvador rivers or through car-avan movements toward the coast. The logisticalimportance of the site suggests its economic rolebut also speaks to the social relations it may havebeen promoting and engendering.

Conclusion

Evidence from this small sample from San Sal-vador Cemetery supports the idea that this was apopulation engaged in resource intensification andtrade during the Middle Formative period. Peopleliving at San Salvador, and likely at similar intern-odal locations, were engaged in interactions thatspread across the breadth of the Atacama Desertfrom the coast to the Andes. The evidence from SanSalvador suggests that the networks of exchangeand mobility in the Atacama during the Formativeperiod covered large expanses of desert and likelyinvolved people living in isolated settlements ininternodal spaces— spaces that are not, in fact,empty. Of particular note, the dietary variability evi-denced in our stable isotope analyses supports thecontention that the San Salvador population wasinvolved in large-scale exchange networks and mayhave functioned as a logistical outpost. It is note-worthy that the amount of variation seen in thesefew individuals from San Salvador is comparableto that seen in entire regions (e.g., central westernArgentina [Gil et al. 2006]). In these data, we maybe seeing a facet of what Upham has referred to as“‘economic layering,’ a feature of the panregionaleconomic system that linked groups at widely vary-ing levels of complexity over a vast area”(1992:143). It appears that the individuals at SanSalvador engaged in these exchange networks withconsiderable variability, likely resulting in differ-ent dietary patterns. Moreover, these data may infact reflect the varied modalities of human move-

ment in the Formative.Building on Lazzari (2005a), it seems that

exchange did not occur in a simple unidirectionalmanner but rather that varied patterns in the move-ment of goods reflect the different materials beingexchanged and the various social groups involvedin the interactions. Archaeological evidence fromthe Formative period in northern Chile indicatesthat coastal and inland populations were produc-ing surplus economies and a network of exchangefor foods, prestige goods, and ideas at this time(Gallardo 2009). As Jochim notes, “Food exchangehas been viewed as intimately involved in the devel-opment of hunter-gatherer complexity” (2006:80).Interaction between these different groups, there-fore, cannot necessarily be interpreted as a simpletrajectory from a center to a periphery, nor is thiscomplex movement limited to complex polities but,rather, reflects the needs and desires of groups andindividuals regardless of scale as well as the laborand varied social processes they produce. Lazzariwrites, “Circulating things and people refer toplaces and people that are not immediately presentin everyday life, thus concretizing in material formthe presence of other people and places”(2005b:131). It would be worth exploring othersites from this time in order to determine if this pat-tern of dietary variability was unique to these out-posts or part of the larger shifts seen in northernChile at this time. These networks in the Forma-tive period were not circumstantial but were partof a regional economic structure that remains onlypartly explored. Further excavation of similar smalltrade outposts and areas of resource intensificationmay help to flesh out what appear to be increas-ingly complex patterns of interaction. In northernChile, the vast expanses of desert between largesites may suggest swaths of vacant land; however,they provide a prime opportunity for exploringhuman movement, interactions, and the use ofempty or internodal spaces in prehistory.

Acknowledgments. This research was conducted under theauspices of Fondo Nacional de Ciencia y Tecnología#1070083 (F. Gallardo, principal investigator). We alsogratefully acknowledge the Gaylord Fund for PacificResearch at Colorado College. José Blanco, BernarditaBrancoli, Milagros de Ugarte Wilfredo Faúndes, JosefinaGallardo, Pablo Gallardo, Rafael Labarca, GonzaloPimentel, Charles Rees, and Carole Sinclaire all played sig-nificant roles in the excavation of the San Salvador sites andanalyses of the recovered material; their contribution is


greatly appreciated. Dr. Mark Hubbe and the staff of theInstituto de Investigaciones Arqueológicas y Museo Le Paigein San Pedro de Atacama facilitated the skeletal analyses. Inthe Stable Isotope Laboratory of the Department of Earthand Environmental Science at the University of Illinois atChicago, Drs. Miguel Gonzalez-Meler, Javier Lugo-Peréz,and Nuria Gómez-Casanovas, as well as Manette Sandor,graciously provided laboratory access and invaluable assis-tance. Finally, the thoughtful comments provided by the edi-tors and anonymous reviewers improved the article.

References Cited

Agüero, M. Carolina2005 Aproximación al asentamiento temprano en los oasis

de San Pedro de Atacama. Estudios Atacameños 30:29–60.Agüero, M. Carolina, Patricia Ayala, Mauricio Uribe, Carlos

Carrasco, and Bárbara Cases2004 El Período Formativo desde Quillagua. In Esferas de

interaccion prehistóricas y fronteras nacionales moder-nas, edited by Heather Lechtman, pp. 73–120. Institute ofAndean Research, New York.

Ambrose, Stanley H.1990 Preparation and Characterization of Bone and Tooth

Collagen for Isotopic Analysis. Journal of ArchaeologicalScience 17:431–451.

1991 Effects of Diet, Climate and Physiology on NitrogenIsotope Abundances in Terrestrial Foodwebs. Journal ofArchaeological Science 18:293–317.

1993 Isotopic Analysis of Paleodiets: Methodological andInterpretive Considerations. In Investigation of AncientHuman Tissue Chemical Analyses in Anthropology, editedby Mary K. Sanford, pp. 59–130. Gordon and Breach,Langhorne, Pennsylvania.

2007a Bone/Tooth Collagen Preparation. University of Illi-nois at Urbana-Champaign, Anthropology DepartmentEnvironmental Isotope Paleobiogeochemistry Laboratory,Urbana-Champaign.

2007b Sample Preparation for Bone and Tooth CarbonateAnalysis. University of Illinois at Urbana-Champaign,Anthropology Department Environmental Isotope Paleo-biogeochemistry Laboratory, Urbana-Champaign.

Ambrose, Stanley H., and Michael J. DeNiro1987 Bone Nitrogen Isotope Composition and Climate.

Nature 325:201.Ambrose, Stanley H., and Lynette Norr

1993 Experimental Evidence for the Relationship of theCarbon Isotope Ratios of Whole Diet and Dietary Proteinto Those of Bone Collagen and Carbonate. In PrehistoricHuman Bone, edited by Joseph B. Lambert and GiselaGrupe, pp. 1–37. Springer-Verlag, New York.

Ambrose, Stanley H., Jane E. Buikstra, and Harold W. Krueger2003 Status and Gender Differences in Diet at Mound 72,

Cahokia, Revealed by Isotopic Analysis of Bone. Journalof Anthropological Archaeology 22:217–226.

Baldini, Lidia2007 Cancha de Paleta, un cementerio del Periodo Forma-

tivo en Cachi (Valle Calchaqui, Salta). Cuadernos de laFacultad de Humanidades y Ciencias Sociales 32:13–33.

Ballester, Benjamin, and Francisco Gallardo2011 Prehistoric and Historic Networks on the Atacama

Desert Coast (Northern Chile). Antiquity 85(329):875–889.Berenguer, José

2004 Caravanas, interacción y cambio en el desierto de Ata-cama. Sirawi Ediciones, Santiago.

Bösl, Cornelia, Gisela Grupe, and Joris Peters2006 A Late Neolithic Vertebrate Food Web Based on Sta-

ble Isotope Analyses. International Journal of Osteoar-chaeology 16:296–315.

Buikstra, Jane E., and Douglas H. Ubelaker1994 Standards for Data Collection from Human Skeletal

Remains. Arkansas Archeological Survey, Fayetteville.Buzon, Michele R., Jacqueline T. Eng, Patricia M. Lambert, and

Phillip L. Walker2005 Bioarchaeological Methods. In Handbook of Archae-

ological Methods, edited by Herb A. Maschner andChristopher A. Chippendale, pp. 871–918. AltaMira Press,Lanham, Maryland.

Calvin, Melvin, and Andrew A. Benson1948 The Path of Carbon in Photosynthesis. Science

107:476–480.Cases, Bárbara, Charles Rees, Gonzalo Pimentel, Rafael

Labarca, and Daniela Leiva2008 Sugerencias desde un contexto funeraria en un “espa-

cio vacío” del desierto de Atacama. Boletín del MuseoChileno de Arte Precolumbino 13(1):51–70.

Chisholm, Brian S., D. Earle Nelson, and Henry P. Schwarcz1982 Stable-Carbon Isotope Ratios as a Measure of Marine

Versus Terrestrial Protein in Ancient Diets. Science216(4550):1131–1132.

Craig, Harmon1957 Isotopic Standards for Carbon and Oxygen and Cor-

rection Factors for Mass-Spectrometric Analysis of Car-bon Dioxide. Geochimica et Chosmochimica Acta12:133–149.

Dembo, Adolfo, and José Imbelloni1938 Deformaciones intencionales del cuerpo humano de

carácter étnico. J. Anesi, Buenos Aires.DeNiro, Michael J., and Christine Hastorf

1985 Alteration of 15N/14N and 13C/12C Ratios of PlantMatter During the Initial Stages of Diagenesis: StudiesUsing Archaeological Specimens from Peru. Geochimicaet Cosmochimica Acta 49:97–115.

de Souza, Patricio, Isabel Cartajena, Lautaro Núñez, and Car-los Carrasco

2010 Cazadores recolectores del Arcaico Tardío y desarrollode complejidad social en la Puna de Atacama: las eviden-cias del sitio Tulán-52 (norte árido de Chile). Werken13:91–118.

de Ugarte, Milagros, and Francisco Gallardo2010 Explotación del recurso forestal y vegetal en el Río

San Salvador durante el Formativo Medio, desierto deAtacama (norte de Chile). Report submitted to FondoNacional de Ciencia y Tecnología, Project 1070083. San-tiago.

Dillehay, Tom D., Verónica I. Williams, and Calogero M. San-toro

2006 Areas periféricas y nucleares. Contextos de interac-ciones sociales complejas y multidireccionales. Chungara38:249–256.

Falabella, Fernanda, M. Teresa Planella, Eugenio Aspillaga,Lorena Sanhueza, and Robert H. Tykot

2007 Dieta en sociedades Alfareras de Chile Central: aportede análisis de isótopos estables. Chungara 39(1):5–27.

Finucane, Brian, Patricia Maita Agurto, and William H. Isbell2006 Human and Animal Diet at Conchopata, Peru: Stable

Isotope Evidence for Maize Agriculture and Animal Man-agement Practices During the Middle Horizon. Journal ofArchaeological Science 33:1766–1776.

Fuller, Benjamin T., James L. Fuller, David A. Harris, and RobertE. M. Hedges

2006 Detection of Breastfeeding and Weaning in Modern


Human Infants with Carbon and Nitrogen Stable IsotopeRatios. American Journal of Physical Anthropology129:279–293.

Gallardo, Francisco2009 Social Interaction and Rock Art Styles in the Ata-

cama. Antiquity 83:619–633.Gallardo, Francisco, Luis Cornejo, Rodrigo Sánchez, Bárbara

Cases, Alvaro Román, and Angel Deza1993 Una aproximación a la cronología y el asentamiento

en el Oasis de Quillagua (Río Loa, II Región). Actas delXII Congreso Nacional de Arqueología Chilena 1:41–60.Temuco.

Gil, Adolfo F., Robert H. Tykot, Gustavo Neme, and Nicole R.Shelnut

2006 Maize on the Frontier: Isotopic and MacrobotanicalData from Central-Western Argentina. In Histories ofMaize, edited by John E. Staller, Robert H. Tykot, andBruce Benz, pp. 199–214. Academic Press, New York.

González, Carlos, and Catherine Westfall2006 Cementerio Regimiento Chorrillos de Calama: testi-

monios funerarios formativos en el Loa Medio, Región deAntofagasta. Actas del XVII Congreso Nacional de Arque-ología Chilena: 17–18. Valdivia.

Hatch, Marshall D., and Charles Roger Slack1966 Photosynthesis by Sugarcane Leaves. A New Car-

boxylation Reaction and the Pathway of Sugar Formation.Biochemical Journal 101:103–111.

Hawkey, Diane E., and Charles F. Merbs1995 Activity-Induced Musculoskeletal Stress Markers

(MSM) and Subsistence Strategy Changes Among AncientHudson Bay Eskimos. International Journal of Osteoar-chaeology 5:324–338.

Horn, Peter, Stefan Hölzl, Susanne Rummel, Göran Åberg,Solveig Schiegel, Daniela Biermann, Ulrich Struck, andAndreas Rossman

2009 Humans and Camelids in River Oases of the Ica-Palpa-Nazca Region in Pre-Hispanic Times— Insightsfrom H-C-N-O-S-Sr Isotope Signatures. In New Tech-nologies for Archaeology, edited by Markus Reindel andGünther A. Wagner, pp. 173–192. Springer-Verlag, Berlin.

Hückstädt, Luis A., C. P. Rojas, and Tarsicio Antezana2007 Stable Isotope Analysis Reveals Pelagic Foraging by

the Southern Sea Lion in Central Chile. Journal of Exper-imental Marine Biology and Ecology 347(1–2):123–133.

Jochim, Michael2006 The Implications of Inter-Group Food Exchange for

Hunter-Gatherer Affluence and Complexity. In BeyondAffluent Foragers, edited by Colin Grier, Jangsuk Kim, andJunzo Uchiyama, pp. 80–89. Proceedings of the 9th Inter-national Conference on Archaeozoology, Durham, Eng-land, 2002. Oxbow Books, Oxford.

Judd, Margaret2008 The Parry Problem. Journal of Archaeological Sci-

ence 35:1658–1666.Kellner, Corina M., and Margaret J. Schoeninger

2007 A Simple Carbon Isotope Model for ReconstructingPrehistoric Human Diet. American Journal of PhysicalAnthropology 133:1112–1127.

2008 Wari’s Imperial Influence on Local Nasca Diet: TheStable Isotope Evidence. Journal of AnthropologicalArchaeology 27:226–243.

Koch, Paul L., Noreen Tuross, and Marilyn L. Fogel1997 The Effects of Sample Treatment and Diagenesis on

the Isotopic Integrity of Carbonate in Biogenic Hydroxy-lapatite. Journal of Archaeological Science 24:417–429.

Kortschak, Hugo P., Constance E. Hart, and George O. Burr1965 Carbon Dioxide Fixation in Sugarcane Leaves. Plant

Physiology 40:209–213.Lazzari, Marisa

2005a Traveling Objects and Spatial Images: ExchangeRelationships and the Production of Social Space. InGlobal Archaeological Theory: Contextual Voices andContemporary Thoughts, edited by Pedro P. Funari, AndrésZarankin, and Emily M. Stovel, pp. 191–210. Springer,New York.

2005b The Texture of Things: Objects, People, and Land-scape in Northwest Argentina (First Millennium A.D.). InArchaeologies of Materiality, edited by Lynn Meskell, pp.126–161. Blackwell, Oxford.

Lazzari, Marisa, Lucas Pereyra Domingorena, Maria CristinaScattolin, Leslie Cecil, Michael D. Glascock, and RobertJ. Speakman

2009 Ancient Social Landscapes of NorthwesternArgentina: Preliminary Results of an Integrated Approachto Obsidian and Ceramic Provenance. Journal of Archae-ological Science 36:1955–1964.

Lee-Thorp, Julia A.1989 Stable Carbon Isotopes in Deep Time: The Diets of

Fossil Fauna and Hominids. Unpublished Ph.D. disserta-tion, University of Cape Town, Cape Town.

Longin, R.1971 New Method of Collagen Extraction for Radiocarbon

Dating. Nature 230:241–242.Lumbreras, Luis G.

2006 Un formativo sin cerámica y cerámica preformativa.Estudios Atacameños 32:11–34.

McCormac, F. Gerry, Alan G. Hogg, Paul G. Blackwell, CaitlinE. Buck, Thomas F. G. Higham, and Paula J. Reimer

2004 SHCal04 Southern Hemisphere Calibration 0–11.0 calkyr BP. Radiocarbon 46:1087–1092.

Malpass, Michael1993 Provincial Inca: Archaeological and Ethnohistorical

Identification of the Impact of the Inca State. Universityof Iowa Press, Iowa City.

Mariotti, André1983 Atmospheric Nitrogen Is a Reliable Standard for Nat-

ural 15N Abundance Measurements. Nature 303:685–687.Martínez, José Luis

1998 Pueblos del Chañar y el Algarrobo: los atacameñosen el siglo XVII. Dirección de Bibliotecas, Archivos yMuseos, Santiago.

Minagawa, Masao, and Eitaro Wada1984 Stepwise Enrichment of 15N Along Food Chains: Fur-

ther Evidence on the Relation Between "15N and Animalage. Geochimica et Cosmochimica Acta 48(5):1135–1140.

Moragas, Cora1982 Túmulos funerarios en la costa sur de Tocopilla

(Cobija), II Región. Chungara 9:152–173.Murra, John V.

1972 El “Control Vertical” de un máximo de pisos ecológi-cos en la economía de las sociedades andinas. In Visita dela Provincia de Leon de Huanuco en 1562, edited by JohnV. Murra, Vol. 2:429–476. Universidad Nacional HermilioValdizan, Huánuco, Peru.

1985 “El Archipielago Vertical” Revisited. In Andean Ecol-ogy and Civilization: An Interdisciplinary Perspective onAndean Ecological Complementarity, edited by ShozoMasuda, Izumi Shimada, and Craig Morris, pp. 3–14. Uni-versity of Tokyo Press, Tokyo.

Nielsen, Axel E.2004 Estudios internodales e interacción interregional en

los Andes Circumpuneños. In Esferas de interacción pre-


históricas y fronteras nacionales modernas en los AndesSur Centrales, edited by Heather Lechtman, pp. 29–62.Instituto de Estudios Andinos, Lima.

Niinimäki, Sirpa2011 What Do Muscle Marker Ruggedness Scores Actu-

ally Tell Us? International Journal of Osteoarchaeology5:324–338.

Núñez, Lautaro, and Tom D. Dillehay1995 Movilidad giratoria, armonía social y desarrollo en

los Andes Meridionales: patrones de tráfico e interaccióneconómica (Ensayo). Universidad del Norte, Antofagasta.

Núñez, Lautaro, Isabel Cartajena, Carlos Carrasco, Patricio deSouza, and Martin Grosjean

2006 Emergencia de comunidades pastoralistas formativasen la Puna de Atacama. Estudios Atacameños 32:93–117.

Núñez, Lautaro, Isabel Cartajena, Carlos Carrasco, and Patri-cio de Souza

2006 El Templete Tulán de la Puna de Atacama: emergen-cia de complejidad ritual durante el Formativo Temprano(norte de Chile). Latin American Antiquity 17(4):445–473.

Pimentel, Gonzalo, Charles Rees, Patricio de Souza, and Patri-cia Ayala

2010 Estrategias de movilidad del Período Formativo en laDepresión Intermedia, desierto de Atacama. In Actas delXVII Congreso Nacional de Arqueología Chilena, Vol.2:1353–1364. Sociedad Chilena de Arqueología, Univer-sidad Austral de Chile. Ediciones Kultrún, Valdivia.

Pollard, Gordon C.1970 The Cultural Ecology of Ceramic-Stage Settlement in

the Atacama Desert. Ph.D. dissertation, Columbia Uni-versity. University Microfilms, Ann Arbor.

Ranson, Stanley L., and Meirion Thomas1960 Crassulacean Acid Metabolism. Annual Review of

Plant Physiology 11(1):81–110.Richards, Michael P., Simon Mays, and Benjamin T. Fuller

2002 Stable Carbon and Nitrogen Isotope Values of Boneand Teeth Reflect Weaning Age at the Medieval WharramPercy Site, Yorkshire, UK. American Journal of PhysicalAnthropology 119:205–210.

Schoeninger, Margaret J., and Michael J. DeNiro1984 Nitrogen and Carbon Isotopic Composition of Bone

Collagen from Marine and Terrestrial Animals. Geochim-ica et Cosmochimica Acta 48(4):625–639.

Schoeninger, Margaret J., Michael J. DeNiro, and HernrikTauber

1983 Stable Nitrogen Isotope Ratios of Bone CollagenReflect Marine and Terrestrial Components of PrehistoricHuman Diet. Science 220(4604):1381–1383.

Schortman, Edward M., and Patricia A. Urban1992 The Place of Interaction Studies in Archaeological

Thought. In Resources, Power, and Interregional Interac-tion, edited by Edward M. Schortman and Patricia A.Urban, pp. 3–15. Plenum Press, New York.

Schutkowski, Holger, Bernd Herrmann, Felicitas Wiedemann,Herve Bocherens, and Gisela Grupe

1999 Diet, Status and Decomposition at Weingarten: TraceElement and Isotope Analyses on Early Mediaeval Skele-tal Material. Journal of Archaeological Science26(6):675–685.

Schwarcz, Henry P., Tosha L. Dupras, and Scott I. Fairgrieve1999 15N Enrichment in the Sahara: In Search of a Global

Relationship. Journal of Archaeological Science26(6):629–636.

Sinclaire, Carole2004 Prehistoria del Periodo Formativo en la Cuenca Alta

del Río Salado (Región del Loa Superior). Chungara 36(Suppl.): 619–639.

Spahni, Jean-Christian1964 Momie Atacamenienne mutilée du Rio San Salvador

(Chili). Bulletin de la Société Suisse des Américanistes28:9–12.

1967 Recherches archéologiques a l’embouchure du RioLoa. Bulletin Société des Americanistes 61:179–239.

Stuiver, Minze, Paula J. Reimer, and Ron W. Reimer2005 CALIB 6.0. Electronic document,

http://calib.qub.ac.uk/calib/, accessed December 12, 2010.Thomas, Carlos, Antonia Benavente, Isabel Cartagena, and

George Serracino1995 Topater, un cementerio temprano. Hombre y Desierto

9:159–170.Tieszen, Larry L., and Michael Chapman

1992 Carbon and Nitrogen Isotopic Status of the MajorMarine and Terrestrial Resources in the Atacama. In Pro-ceedings of the 1st World Congress on Mummy Studies,pp. 409–425. Museo Arqueológico y Etnográfico, Tener-ife.

Tieszen, Larry L., Elysha Iversen, and Steven Matzner1992 Dietary Reconstructions Based on Carbon, Nitrogen,

and Sulfur Stable Isotopes in the Atacama Desert, North-ern Chile. In Proceedings of the 1st World Congress onMummy Studies, pp. 427–441. Museo Arqueológico yEtnográfico, Tenerife.

Torres-Rouff, Christina2003 Shaping Identity: Cranial Vault Modification in the

Pre-Columbian Andes. Unpublished Ph.D. dissertation,Department of Anthropology, University of California,Santa Barbara.

2007 La deformación craneana en San Pedro de Atacama.Estudios Atacameños 33:25–38.

Trombold Charles D.1991 Ancient Road Networks and Settlement Hierarchies

in the New World. New Directions in Archaeology. Cam-bridge University Press, New York.

Trueman, Clive N., Karen L. Privat, and Judith Field2008 Why Do Crystallinity Values Fail to Predict the Extent

of Diagenetic Alteration of Bone Mineral? Palaeogeog-raphy, Palaeoclimatology, Palaeoecology 266(3–4):160–167.

Upham, Steadman1992 Interaction and Isolation: The Empty Spaces in Pan-

regional Political and Economic Systems. In Resources,Power, and Interregional Interaction, edited by Edward M.Schortman and Patricia A. Urban, pp. 139–152. PlenumPress, New York.

van Klinken, Gert J.1999 Bone Collagen Quality Indicators for Palaeodietary

and Radiocarbon Measurements. Journal of Archaeolog-ical Science 26(6):687–695.

Notes

1. The extensive natural mummification of individual 5.1did not allow for detailed analysis of all of the postcranialremains.

2. The delta (") notation represents the ratio of the heav-ier to the lighter isotope "13C to 12C in the case of carbon, 15Nto 14N in the case of nitrogen) in parts per thousand (per millor ‰) relative to an international environmental standard, PeeDee Belemnite (Craig 1957) for carbon and AIR (Mariotti1983) for nitrogen.

3. However, as judged against other terrestrial flora andfauna elsewhere in the world, the #15N signatures of these


plants and animals are somewhat, but not dramatically, ele-vated as a result of aridity effects (see Ambrose 1991;Schwarcz et al. 1999).

4. The formula used is %C4 = ([–25‰ – ("13Cap –9.4‰)]/14‰) x 100 (Ambrose et al. 2003:221).

5. It is not absolutely clear why the "13Cap value obtainedfor the enamel apatite of individual 8.1 is so depleted(–4.4‰) vs. the bone apatite of the same individual, as theyshould, in theory, represent roughly the same adolescent timerange. It seems most likely that this difference is the result ofan idiosyncratic or individual difference in time averagingdue to the duration of formation (i.e., the enamel apatite rep-resents a three-year dietary average, while the bone apatiterepresents 10 years or more).

6. The formula used is %Marine = ([6.7‰ – ("15Nco –3‰)]/[18.4‰ – 6.7‰]) x 100 (Ambrose 1993:83). Whilefreshwater fish in certain ecosystems present enriched "15Nvalues that can confound such calculations, the only "15Nvalue that we possess (Orestias spp. [Tieszen et al. 1992]) islow enough (8.1‰, within the range of terrestrial fauna andC4/CAM plants) as to be unproblematic in this regard.

7. This is directly evidenced by the "13Cap values fromConchopata (Finucane et al 2006).

Submitted: December 21, 2010; Revised: November 11,2011; Accepted: January 24, 2012.


Documents

eating fish in the driest desert in the world: osteological and