Society for American Archaeology

War, Chronology, and Causality in the Titicaca Basin Author(s): Elizabeth Arkush Reviewed work(s): Source: Latin American Antiquity, Vol. 19, No. 4 (Dec., 2008), pp. 339-373 Published by: Society for American Archaeology Stable URL: http://www.jstor.org/stable/25478239 . Accessed: 29/03/2012 12:32
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Society for American Archaeology is collaborating with JSTOR to digitize, preserve and extend access to Latin American Antiquity.

http://www.jstor.org

WAR, CHRONOLOGY, AND CAUSALITY IN THE TITICACA BASIN

Elizabeth Arkush

Period resulted directly from the of theMiddle Horizon polities ofWari and Tiwanaku around A.D. 1000, radiocarbon dates presented here from collapse occupation and wall-building events at pukaras in the northern Titicaca Basin indicate these hillforts did not become com mon until late in the Late Intermediate Period, after approximately A.D. 1300. Alternative explanations for this late esca lation ofwarfare are evaluated, especially climate change. On a local scale, the shifting nature ofpukara occupation indicates cycles of defense, abandonment, reoccupation, and wall building within a broader context of elevated hostilities that lasted for the rest of the Late Intermediate Period and beyond. En el Periodo ter Intermedio Tardio (ca. 1000-1450 d.C), los habitantes de muchas partes de la sierra andina abandonaron renos productivos para asentarse en sitios defensivos en colinas, a veces construyendo asentamientos amurallados en las cum " tanto en Quechua como Aymara. Este cambio demuestra una preocupacion por la guerra no conocida bres, llamados "pukaras

In theLate Intermediate Period (ca. A.D. 1000-1450), people inmany parts of theAndean agricultural lands to settle in defensive sites high on hills and ridges, frequently building Quechua and Aymara. This settlement shift indicates a concern with warfare not equaled logical sequence. While the traditional assumption is thatwarfare in the Late Intermediate

highlands moved away from rich hilltop forts known as pukaras in at any other time in the archaeo

internificacion tardia de la guerra, sobre todo las condiciones ambientales adversas. En una escala mas pequeha, los ocupa ciones variables de pukaras indican ciclos locales de la defensa, del abandono, de la reoccupacion, y de la construccion de murallas dentro de un contexto mas amplio de conflicto agravado que duro el resto del Intermedio Tardio y aun despues.

anteriormente en la secuencia arqueologica. Segun la interpretation tradicional, el conflicto resulto directamente del colapso de Wari y de Tiwanaku, aldredor de 1000 d.C, pew los fechados radiocarbonicos de un grupo de pukaras de la cuenca septen trional del Titicaca, presentados en este trabajo, indican que la mayoria de estos pukaras nofueron construidos ni ocupados 1300 d.C. Se evalua las explicaciones alternativas para esta hasta el final del Intermedio Tardio, despues de aproximadamente

The

Late Intermediate Period, ca. A.D. 1000-1450, has long been seen by archae ologists as an era of conflict and political fragmentation (Julien 1993;Matos 1999; Parsons and Hastings 1988). Many parts of theAndean highlands witnessed a settlement shift to more
abounded, and protected with massive (e.g., stone-built walls

lamented nearly two decades ago, All too oftenwe have been forced to talk in
static terms about must a period nearly 500 years critical of decay of the of devel organi long which information inmajor Middle opment ultimately cultural provide

regarding and

the dynamics the dynamics pan-Andean

defensively located sites.Hilltop fortsand refuges

sometimes ditches D'Altroy and Has

inter-regional Horizon for Late Horizon

systems

torf2001; Hyslop 1976; Parsons et al. 2000). The outlines of this pattern have been recognized for some time,particularly for the central and south ern Andean highlands, and recent research has confirmed it inother areas (see below). However,
a better understanding of the causes and ramifica

zation [Parsons and Hastings 1988:228].

Even fine-grained now, most ceramic studies, sequences unable

to construct

or run numerous

radiocarbon dates, have littlechoice but to treatthe Late Intermediate Period as a monolithic and ological vision of a four-century-long epoch of
continuous ods of peace, warfare, in which we cannot see short-term political consolidation, peri or homogenous era. The result is an artificial archae

Period chronologies. As Parsons and Hastings

Elizabeth Arkush Department (arkush@virginia.edu)

tions of thiswidespread conflict has been ham pered by the lack of refined Late Intermediate

of Anthropology, University of Virginia, P.O. Box 400120, Charlottesville, Latin American Antiquity 19(4), 2008, pp. 339-373 by the Society forAmerican Archaeology Copyright ?2008 339

VA 22904

340LATIN AMERICAN

ANTIQUITY sons that these

[Vol. leaders deem

19, No. 4, 2008 sufficient: to accrue

Late IntermediatePeriod, includingpop within the ulation growth, settlement shifts, political consol warfare (Covey 2008; Earle and intensified idation, etal. 1980,1987; Julien 1988;Nielsen 2002; Owen 1995; Stanish 1985, 1992, 2003). This paper pre sentsnew evidence on thechronology ofLate Inter mediate Period warfare from one of the areas in was firstrecognized: the Titicaca Basin of which it southernPeru. Here, a suite of radiocarbon dates brings intobetter focus theprocesses of wall con

the shiftingconstruction and abandonment of set tlements.Yet where more detailed chronological data are available, they indicate significantchange

wealth and greater power, buttress legitimacy, deflate a rival faction,or perhaps even fulfillan ide ological imperative. It ismore difficult to explain warfare where decision-making power is diffused have a throughout society and individual fighters whether and how to fight. Such good deal of say in
contexts span the gamut from decentralized, egal

coercive power. Given the obvious disincentives, and abandonment, struction, defensive why do people with some latitudeforchoice choose occupation, on the toengage incollective violence, andwhy have they reuse of fortifiedsites.These data shed light causes ofwarfare on bothmacroregional and local done so with such appalling frequency?
scales.

itarian societies toweakly centralized chiefdoms, and extend to contemporary factional conflicts in which individual fighters may be galvanized by influential figures or pettywarlords with limited

Explaining Warfare: Scales ofAnalysis and Violent Cycles When discussing group violence across a spectrum of regions and centuries, it is useful todefine "war fare"broadly: as a stateof hostilitybetween groups of people who consider themselves separate polit ical communities, whose members engage in armed, potentially lethal,culturally sanctioned vio
lence and acts of destruction against one another.

ber of levels (see Allen and Arkush 2006, Fergu son 1984, 1990, 2001; Snyder 2002), and the reflects both theoreticalcomplexity of this literature of perspectives thatflourish inanthro thediversity pology and the truly tangled web of conditions,
and violence as rewards?not in human to mention societies. Some pretexts?for scholars motivations,

Anthropologists

have

sought

causes

on a num

view warfare

This definition borrows from Ferguson (1984), Webster (1998), Meggitt (1977),Milner (1999), and Warfare and follows currentusage inanthropology. violence thusdefined excludes acts of interpersonal between individuals, as well as collective violence directed specifically at one individual, such as an

sal drives rooted in theevolution of our species: ter ritorial expansion and defense (Thayer 2004; Wrangham and Peterson 1996), male competition for mating opportunities (Chagnon 1988;Daly and Wilson 1988), xenophobia based in kin selection (Shaw andWong 1989), or a combination of such urges (Gat 2000a, 2000b; van der Dennan 1995).
Materialist as explanations, response by contrast, to population view warfare pressure, a contingent

the expression

of univer

crisis (e.g., Fer execution (Kelly 2000). It also excludes strictly resource stress, or environmental Harris 1974; Vayda 1976). contained "ritual battles" such as Andean tinku, guson 1990, 2001; which do not arise from hostile intergroup rela While applicable tomodern-day conflicts (Gled a great range itsch 1997; Homer-Dixon 1999) and historically it encompasses However, tionships. materialist of group violence with significantdifferences in recorded wars (Zhang et al. 2007), the found favoramong view ofwarfare has particularly from occasional, and consequences, conduct expe

dient killings and ambushes between enemy vil

to the organized, well-supplied, long-range

between periods ofwarfare and episodes of nutri lages, tional stress,rising populations, or environmental campaigns of centralized polities involving thou crisis (Bamforth 2006; Billman et al. 2000; Haas sands of soldiers. The causes ofwar, too, may differsignificantly, 1999; Haas and Creamer 1993; Jones et al. 1999; Lambert 1997,2002; LeBlanc 1999,2003; Lekson depending partly on the degree towhich warring societies are politically centralized (Ferguson 1984, 2002; Milner 2007; Nunn et al. 2007; Petersen 1988; Raab and Larson 1997). This view also draws 1990; Keeley 1996; Reyna 1994). Where small some support fromethnology: for instance,Ember leaders can order large num of powerful numbers and Ember (1992) findthatacross cultures,frequent for rea into battle, wars are bers of soldiers
pursued

archaeologists,

who

draw

robust

connections

Arkush]

WAR, CHRONOLOGY,

AND CAUSALITY

INTHE TITICACA

BASIN

341

material and materialists may assign blame to the

political rewards of warmongering for warriors,

warfare is correlatedwith thefearof unpredictable natural disasters affectingresources.Alternatively,

framework that informs their decisions or long termtrends in material scarcity thatencourage pat terns heightenedwarfare;more eternal still is the of
evolved human psychology that makes war possi

ble. Yet anthropological debates over thecauses of leaders, and aggrandizers (Ferguson 1990). that war (for instance, Gat 2000a, 2000b; Ferguson Meanwhile, political explanations argue war arises when is not possi stable peacemaking 2001) have often been clouded by incompatible
ble because

(Chagnon 1968) to conflict-ridden "failed states" such as Somalia and Yugoslavia (Brubaker and Laitin 1998; Desjarlais and Klein man 1994; Jackson 1990; Simons 1995). Anthro Yanomamo

(1968) thatwarfare is a normal state for peoples This without supra-local governmental institutions. to cases from the explanation has been applied

Hobbesian viewpoint that nonexistent, a distinctly echoes the stance of Sahlins (1968) and Service

authority

structures

are

too weak

or

scales

processes operating in the longueduree are distinct fromshorter-term processes, though theyarticulate and likewise, thatlargerregional scales, with them, especially those that encompass politically inde pendent communities, are more likely to reflect
underlying tic and rather

of analysis.

Here

I assume

that causes

and

Some attempts have been made atmore holis

comprehensive theories of war's causes.

than proximate

causes

of war.

For instance,Ferguson (1990) proposes amodel in have also examined correlations between which a nested hierarchy of determinants? pologists warfare and aspects structure of social infrastructural(population, resource availability, (e.g., Otter bein 1970, 2004). Sahlins (1961) links territorial subsistence technology), structural (kinship and conflict to segmentary lineage organization, which sociopolitical organization), and superstructural allows politically autonomous segments to band and constrain (belief, ''culture")?progressively lines of genealogical relatedness. Kelly (2000) extends this hypothesis, concluding that in seg
together against enemies in ever-larger forces along

influence thepractice ofwar.While a great step in the rightdirection, this model fails to capture the
repeated wars themselves alter material and social con struc cultural understandings,

way

mentary

societies,

an attack

on one

isperceived, not as an individual offense,but as an attack on thewhole group by a whole offending that group, and potentially
requiring retaliation against any member of

group member

ditions,

ture (Allen and Arkush

frequent warfare areas,

2006).

For instance,
resource scarcity

itself exacerbates

group,

leading to endless
approaches (e.g.,

through thedisplacement of refugees tomarginal

population nucleation for defense, and the

vendettas.

Meanwhile, and Tumu

cultural

Bonta
Wiessner

1999; Robarchek
1998)

and Robarchek
stress

1998;

creation of unutilized bufferzones (LeBlanc 2006). Warfare may affect social patterns and cultural val
ues, causing people to valorize martial prowess

the understand

ing of violent action within its culturalmatrix, as a contextually specific social event freighted with meaning. For these scholars, individual and group decisions about war are strongly affected by the socialization of children to react to injury, the inter pretation of death and disease as hostile sorcery,
ideals

(Allen 2006), socialize theirchildren to fear out siders (Ember and Ember 1992; Kusimba 2006), and laud elites for military exploits. Snyder (2002)
proposes a complex each rial instead evolving other. that war system and its causes be seen mate norms as inwhich and warfare, cultural

constraints,

social

patterns,

defined ethnic identities,and other elements of cul turalpractice and worldview. It should be clear from thisbrief overview that these causal factorsoperate on differentspatial and temporal scales (some ofwhich aremore amenable to archaeological investigation than others), and thatby and large they articulate with each other, rather thanbeing mutually exclusive. The motiva
tions of warriors

of masculinity,

autonomy

and honor,

starkly

affect

This paper explores thepotential of a multilay ered explanatory approach towarfare for theLate IntermediatePeriod through the examination of a kind of material culture particularly amenable to

shorter more local scale thaneither thecultural and

in a particular

war

exist

on

spread and increased defensibility of fortifiedsites as is interpreted evidence for an elevated threatof war. But. in linewith the above discussion, I also

archaeological analysis: the fortified site. Partly, is fortification used here (as it normally isby archae as an index of war, so that the ologists) simply

342LATIN AMERICAN

ANTIQUITY

[Vol.

19, No. 4, 2008

and defensive settlement propose thatfortifications had thepotential to alter regional sociopo patterns litical landscapes, themselves influencingchoices
about violent action

tlement patterns, fortifications,and buffer zones mark social categories on the land,making allies and enemies more easily inheritable.They allow
their users to plan offensives more securely, encour

for generations.

Defensive

set

sons et al. 1997,2000) andAsto (Lavalee and Julien 1973), suggest small-scale raiding for stores and livestock,while the densely occupied hillforts of the upper Mantaro (D'Altroy and Hastorf 2001; Earle et al. 1980, 1987; Hastorf et al. 1989) indi
cate warfare on a much larger scale. South of

aging aggressive solutions todisputes. Individually, fortifiedsites are difficulttovanquish; in multiples, fortsimpede theconquest and stable controlof hos tile territory forcing conquerors to capture and by garrison each fort. They thus tend toentrenchexist When closely controlled by ing political patterns. a central authority,fortscement thatauthority, but in contexts of fragmentation,heavily fortifiedter rain is especially prone to repeated cycles of incon clusive violence because itis so difficulttoconquer
of landscape, settlement, and walls alter polit

Ayacucho and into Andahuaylas, patternsof defen sive and fortifiedsites continue, varying in scale (Meddens 1999; Valdez et al 1990, 1994; Vivanco 1999). There is some evidence of warfare around the margins of theCuzco area (e.g. Kendall 1996),

although thebirthplace of the Incas features mostly nondefensive settlement in theLate Intermediate Period (Bauer 1992,2004; Bauer and Covey 2002; Hefferman 1996), suggesting that political consol idation proceeded either peacefully or too rapidly to result infortification. dominated the Hilltop forts Titicaca Basin in the Late Intermediate Period (Arkush 2005; Frye and de laVega 2005; Hyslop 1976;Neira 1967; Stanish 2003; Stanish et al 1997).

and consolidate. In these ways, novel defensive

uses

ical relationships for the long term.

The Problem of Warfare in the Late Intermediate Period The warfare of theLate IntermediatePeriod begs
for an adequate explanation because of its sheer

scale. The pattern of defensive hilltop settlements Nielsen 2001). The upper portions of Pacific coastal valleys and fortifiedsites was recognized for the central In northern Peru, of warfare. and southern highlands decades ago, but recent also betray evidence research has demonstrated that it extends from at defensivewall systemsand strategically placed for tifiedsitesof theupper Jequetepeque, Chicama, and least northern highland Peru, throughout central and southernPeru and highland Bolivia, to north Moche watersheds may have controlled highland coast traffic ern Chile and protected Chimor from highland The north and northwestern Argentina. ern Peruvian incursions (Julien 1988; Krzanowski 1977, 1983; and around sierra Cajamarca is dotted with hilltop settlements, Topic and Topic 1987). In thecentral and southern Huamachuco coastal valleys?Chillon (Farfan 1995; Silva 1992), sometimes lightly fortified (Julien 1988, 1993; Topic and Topic 1987), and defensive siting,forti
fication, and weapons such as sling stones and

ized populations built and used the plentiful hill fortsof theBolivian altiplano (Lecoq 1997; Lecoq and Cespedes 1997; Nielsen 2002), northern and eastern Chile (e.g. Llagostera and Costa 1999; Nunez and Dillehay 1978; Schiappacasse et al. 1989) and northwest Argentina (DeMarrais 1997;

Further south, smaller and less politically central

maceheads

are common

in Late

Intermediate

Period sites of theChachapoyas region (Narvaez 1987; Schjellerup 1992, 1997). Patterns of hilltop settlementand some fortificationcharacterize the Callejon de Huaylas and both sides of the upper Maranon River (Bonnier 1978;Mantha 2006;Was silowsky 1999). The central Peruvian highlands around Junfn, Jauja, andAyacucho have benefited
from

Colca (Wernke 2003), Moquegua (Moseley 1989; Owen 1995; Stanish 1992), and valleys to the south (e.g.,Reindel 2005; Santoro et al. 2004)?fortified hilltop settlements suggest endemic warfare with out such centralized political administration. There is also a small but growing body of bioar

chaeological evidence for violent conflict in the Late Intermediate Period (de la Vega et al 2005; Jakobsen et al 1986-1987; Nystrom and Verano 2003; Torres-Rouff and Costa 2006; Verano 2002). Several of these studies find that cranial trauma
common on both males and females. This may

strate systems of hilltop walled settlement.Small, dispersed defensive sites,as inhighland Junfn(Par

archaeological

surveys

that clearly

demon

was

indicate that warfare was not confined to pitched

Arkush]

WAR, CHRONOLOGY,

AND CAUSALITY

INTHE TITICACA

BASIN

343

battles, but included ambushes, raids, and mas

sacres directed at noncombatants?a pattern

The political disturbancesand economic chaos

that followed lapse in the wake reflected of Tiwanaku's col are brutally of in the characteris of this period: the

driving them into remote mountain

across half a continent. Yet while

to theLate Intermediate Period, of course, but at no other timedid itso stronglythreaten populations, fastnesses
was pan warfare

cal of relatively small-scale, decentralized societies (Keeley 1996;Milner 1999). The geographic extent and intensity thiscon of was unprecedented.Warfare was not unique flict

typi

tic pattern

settlement

fortified village.

Andean, individual wars must have been mostly for Chimu local or subregional affairs, outside of the

empire and theCuzco Valley, polities were small in scale. There are only a handful of known sites with over 500 houses, mostly regional centers in the central and south-central highlands. Even in
the areas chies, around they these centers, settlement hierar geo when are present at all, are small

conflicts to thedecay of the strong Weberian state and Laitin 1998; Desjarlais and Klein (Brubaker man 1994; Jackson 1990), and even with theway thepopular press traces them to ancient "tribal" or ethnic hatreds that are somehow unloosed when statesweaken. This is not to critique the standard explanation of Late Intermediate Period violence on theoretical grounds, for it is genuinely plausi ble: state collapse has oftenbeen followed by war

This viewpoint fallswithin the tradition polit of ical explanations of war: thatwar is expectable when not suppressed by governments. Italignswith the way social scientists trace modern-day factional

or graphically, and satellite sites are often fortified defensible, demonstrating that the larger centers could not fullyprotect theirvicinities. Centralized storage facilities are almost nonexistent in thehigh lands, and there is only limited evidence for site planning, indicatinga pervasive decentralization of

Yet the wide extent political and economic activity. of evidence for conflictundermines the idea that it
ultimately resulted from local processes and con

and their disintegration may have caused great social disruption even in regions never under their direct purview (forLate IntermediatePeriod forti fication and defensive settlement patternsoccurred well beyond theareas of former Wari and Tiwanaku control). Instead, the hypothesis is best tested through fine-grainedchronologies: did warfare fol low on theheels of state collapse or not? Alternatively, some archaeologists have turned toenvironmental change toexplain intensified con flictwithin their regions (Nielsen 2001, 2002;

fare.The Middle Horizon polities ushered in new realms, ways of organizing societies over far-flung

ditions. A plausible explanatory frameworkmust apply across the scale of theAndes, while at the same time recognizing the local nature of political agency in this timeof fragmentationand regional
ization.

Explaining Warfare in the Late Intermediate Period Because theLate IntermediatePeriod is by defin ition a hiatus after the Middle Horizon collapse,
warfare has

Seltzer and Hastorf 1990; Torres-Rouff and Costa studies 2006). As more paleoclimatological emerge, itbecomes possible to draw connections
between climate change and major cultural transi

on Yet information prehistoric climate is still tions. and difficultto interpret, and archaeologists partial
must The why examine correlations chroniclers, were too, touched in timing carefully. on the question era. of

cussions have followed his lead. A typical exam ple can be drawn fromKolata (1993:299): The demise of theTiwanaku empire brought with itwidespread political instability. The
"Pax Tiwanaku" no longer repress imposed by the empire could inter-ethnic hos ingrained, the former provinces of the empire polities other bitterly natural contest and resources.

growth of thatcollapse. This position was initially outlined byHyslop (1976:134), andmany otherdis

traditionally

been

seen

as a direct

out

wars

eral statements clouded by Spanish and Inca cul

tural biases, they attribute pre-Inca warfare to

fought

in the pre-Inca

In gen

tilities, and dissolved

into small

stress the material goals of groups and individuals. For instance,descriptions ofwarlords (such as the Colla paramount lord) indicate thatone ostensible goal was theconquest and political controlof larger
territories and

political anarchy and to a vaguely imputed culture of bellicosity. As proximate causes of war, they

ing land, water,

aggressors sought towrest land from their foes,

subject

populations.

However,

some

344

LATIN AMERICAN

ANTIQUITY

[Vol. 19, No. 4, 2008

: -t

(W

, BRAZIL

&_

'

?JIB '::

.
'.
k

peru '~%M" y y \^ima.,.;;;-.'. ] ^ ': ' ^-Y _^BOLIVIA \ ''''''-''fllyflBl area -->#fOI-iillPlgy-; &??M: V ^*?. TiTI "JF^I^^^H^HK -~: shown I ~W~-Z:-rZ:i Jj*"""""
*ljU Tjs"' TlMpjlfi|L~

^jflfflfllffp- "^^^^^^^^^^^^^

0 10 20
Figure rather jects, sure,

40

60

80

J^M4

~'\

'l||||liiHl^^Brj_-i
in the early Colonial

^wll
period.

1. The Lake Titicaca and

Basin, with the distribution them as sub pres were

of ethnic groups

than conquer suggesting and stresses

incorporate

that shortages, on marginal

population groups

were fought over "water and land, which they would take from each other" (Cobo 1979:97 [1653]). Cieza also states thatpre-Inca Andeans took "the spoils that theyfound and thewomen of the vanquished"

implicated: Cieza de Leon asserts thatwars were fought "over the farmlands or for other things" wars (Cieza 1985:6 [1553]), and Cobo concurs that

ties intofollowing them?encouraged and perpet own power, and uated wars inorder to solidify their disproportionately won lands and wives inwar.
Here, causal factors

A settingof political fragmentationallowed poten tial competition between communities. Tensions

over from (viewed land and resource by other resources may have stemmed for women evolved male shortage. some Competition as an scholars

operated

on a number

of scales.

LeBlanc's analysis (1981:340-353) ofToledo's Wanka area elegantly highlights the visita in the interplayof factors in local decisions to go towar.
Informants wars stated over their pre-Inca arable ancestors as fought stored primarily land, as well chosen coerced

1985:6[1553]).

back to their pukaras

(Cieza

behavioral pattern) may have been a significant factor in the supportof individualwarriors forpar
ticular wars. machinations Finally, wars were encouraged by the of ambitious sinchis.

The Colla Region The Titicaca Basin

arena for investigating

in the Titicaca Basin

food,women, and livestock,but theyalso said that

sinchis?local in war, or who war leaders for their prowess their communi sometimes

is a particularly appropriate
the causes of Late Interme

diate Period warfare (Figure 1). Itwas one of the regions where the pattern of Late Intermediate

Arkush]

WAR, CHRONOLOGY,

AND CAUSALITY

INTHE TITICACA

BASIN

345

Period warfare was firstnoted (Hyslop 1976), and notable it is indeed: the steep hills thatjut from the flatplains of the basin are dotted with the impos ing hillforts known as pukaras in Quechua and Aymara. Since theTiticaca Basin housed the state of Tiwanaku, ifLate IntermediatePeriod warfare erupted as a resultof statecollapse, thisarea should

colla (Johnson 2003). North of Paucarcolla, Tiwanaku pottery is present in small quantities (e.g., at themouth of theHuancane valley [Stan ish et al. 2005]), and a largely unrelated ceramic tradition,provisionally termedHuana, may have
been

have been quickly affected. In the Middle Horizon, the southernTiticaca Basin was dominated by Tiwanaku, while the northernbasin was not fully incorporated (Stanish et al. 2005). Tiwanaku centers were established near modern Puno (Schultze 2000) and Paucar

realms (senorios) led by powerful and possibly hereditarywarlords, yet the archaeological land scape of numerous pukaras, relatively small site hierarchies, and rather subtle status distinctions within communities, suggests less unified and less hierarchical societies (Arkush 2005; Frye 1997; Frye and de laVega 2005). However, the chroni
clers' accounts

279 [1553], 1985:15, 22, 110, 121 [1553]; Cobo 1979:139-140 [1653]; Sarmiento 1988:105-106 [1572]). These groups, particularly theCollas and the Lupacas, are described as politically unified

settlement ported by defensive and fortified patterns found throughout most of the Titicaca Basin (Arkush 2005; Barreda 1958; Bennett 1933,1950; Frye 1997; Hyslop 1976;Neira 1967; Stanish etal. 1997; Stanish 2003; Tapia 1978a, 1978b, 1985; Tschopik 1946). This pattern can be seen as a

of frequent

warfare

are

clearly

sup

used by contemporary, non-Tiwanaku affili regionalmanifestation of theverywidespread con ated peoples (Stanish 2003). Because Huana has flictof theLate IntermediatePeriod. to be identified (that is, differentiated While several archaeological studieshave inves only begun from earlier Pucara or laterCollao pottery), Mid tigated theneighboring Lupacas to the south (de la dle Horizon settlementand society in thenorthern Vega 1990; Frye 1997; Frye and de laVega 2005; basin are not well understood. Collas of the Hyslop 1976; Stanish et al. 1997), the can be dated to approxi Tiwanaku's Late Intermediate Period have been surprisingly collapse understudied for a group with such stature in the mately A.D. 1000,when itscolonies in Moquegua were abandoned and monumental construction at ethnohistoric literature. Several nonsystematic the city of Tiwanaku ceased, although some occu reconnaissances and sitevisits (Fuentes 1991;Neira 1967; Palacios 1934; Rowe 1942; Tschopik 1946; pation and craftproductionmay have lingered on in thecityuntil ca.A.D. 1150 (Janusek2004; Owen 1940a, 1940b) clearly established the Vasquez Late IntermediatePeriod, populations characteristicLate IntermediatePeriod settlement 2005). In the in the better-studied southern and southwestern pattern of fortified pukaras, defensible unfortified basin moved raised to dispersed settlements, fields and lake margins abandoning for areas more suited sites, and tombs ranging burial from tower-like to cist chullpas graves. (aboveground structures)

topastoralism and rainfallagriculture (Frye and de laVega 2005; Janusek 2004; Stanish et al 1997). Hilltop settlementbecame common in nearly all Titicaca Basin, including the portions of the pukaras that constitute the focus of this study.The Late IntermediatePeriod also saw changes inburial pat ternsand ceramic styles,and thecessation of long
mounds, sunken courts, and monoliths). Contact

Excavations at Hatuncolla (Julien 1983) revealed thatthepurportedpre-Inca capital of the Collas was
in fact an intrusive Inca settlement,

tion the political centralization of theLate Inter mediate Period Collas, although the long-termuse of Sillustani as a major burial center (Ayca 1995;

opening

to ques

lived forms of ceremonial architecture (platform

Revilla and Uriarte 1985; Ruiz 1973) may indicate a certain cultural or ethnic unity for theCollas, or a subgroup of Collas.

period ethnohistories describe regional polities warfare: the Collas in thenorth engaged infrequent ern basin; theLupacas, neighbors and bitter rivals

Pukaras

on the southwestern side of the lake; the Pacajes south of the lake, inTiwanaku's formerheartland; and the Canas, theCollas' enemies to the north (Betanzos 1996:93 [1551-1557]; Cieza 1984:274,

The data used in thisanalysis come from a project that investigatedpukaras of thenorthernand north western basin in the territory attributedto theCol las (Figure 2). Among other aims, theproject was

346LATIN AMERICAN

ANTIQUITY hi ii 11m_

[Vol.

19, No. 4, 2008

. .ini ^a*]jj^"'^^s^^^^j))j^^^ri5^^

" '' ' i ' ' '" '"" -"'"''*?'.".-~^^T!??j^""T*^jS^

Figure 2. Pukaras

in the northwest Titicaca

Basin.

designed to determine thechronology of fortifica tion in the region and potential correlations in tim ingwith events such as the collapse of Tiwanaku.
Pukaras were identified based

ground-checked were obtained wall mortar

from

and mapped, and carbon samples 1-x-1 -m test pits and defensive of 15 pukaras. Using air pho

at a subset

plemented by ground sighting, toponyms, and the A published literature. sample of 44 pukaras was

on air photos,

sup

tos for the identificationof pukaras permitted an investigation thatwas both systematic and geo graphically extensive. Thus, while Colla pukaras

Arkush]

WAR, CHRONOLOGY,

AND CAUSALITY

INTHE TITICACA Table

BASIN

347

broadly resemble pukaras of theLupaca area (Frye 1997; Hyslop 1976), the large number assessed here allows for a more comprehensive picture of However, this methodology did leave some impor nonfortified Late Intermediate Period settlement. Several recent and ongoing projects of full
coverage tant questions unanswered, such as the nature of variability in pukara use, size, and chronology.

1.Titicaca Basin Chronology. Date Range_ A.D. Phase Phase A.D. Late _Middle II I A.D. A.D. 1450-1532 1300-1450 1000-1300 400-1000 200 B.C.-A.D. 1300-200 B.C. 400

Period_Phase Late Horizon Late Intermediate Period Middle Horizon

Formative

Plourde and de laVega 2001; Stanish et al. 2008)

to elucidate the complete settlement

slingstone fire in the defense of these sites, and indeed,piles of unworked rivercobbles are present pat promise near thewalls at several pukaras.Where there is ternsof theLate IntermediatePeriod. In thebetter studiedLupaca area to the south, small, unfortified more than one defensive wall (in 37 out of 44 pukaras), thewalls are nearly always placed 15 to villages and single-household hamlets cluster near zones between them 30 m apart,well under theeffectiverange of slings pukaras, with possible buffer (at least 50 to 60 m; see Keeley [2007:73]; Rawls (Frye 1997). Colla pukaras of the northernTiticaca Basin
to a basic canon in terms of location, artifacts. are wall conform

pedestrian

survey

in the Colla

zone

(e.g.,

construction, the design and placement of house

structures, and

[1975:130]). This positioning suggests that the spaces between walls were partly intended as "killing alleys," trapping attackers in a restricted
zone where defenders along the inner wall could

higher pasturage. They have excellent views of

their surrounding ber of other above can the range well area, including, Because usually, a num are pukaras. most pukaras

on hills of varying size and steepness between about 3,900 and 4,600 masl, with most about 4,100 masl, and they are preferentially located near the plains and river valleys, giving them access to both hill-base agricultural zones and

typical

They

found

direct a barrage of projectiles at them. The design of entrances, necessary weak points
in walls, also reveals defensive

entrances consist either of several small, linteled

doorways, or a smaller Gates number are often of larger gates, baffled, or a combination. screened

pressures.

Pukara

with a separate parallel wall inside, or flanked by inset walls or platforms to each side, allowing the entrance to be monitored by defenders standing above (see Keeley et al. [2007] for a general treat
Doorways are small (60 cm wide on aver

of cultivation, preserved.

their architecture

ment

The most impressive form of standing archi tecture is the multiple defensive walls that supple
the naturally strategic setting of these sites,

be quite

ment).

encircling the hilltop and habitation areas. These walls were clearly designed to formdefensive bar
riers in a cost-efficient manner.

and could have been age), forcing single-fileentry, easily blocked from the insidewith rubble or brush, especially where they slope uphill through the
walls. In other words, these constructions are

tage of cliffsor rock outcrops, linkingthesenatural defenses intoa chain. They are largeston the sides of the hill that are most approachable, and they
may peter out on steep terrain. Defensive walls are

They

take advan

clearly intended fordefense. The ongoing concern with warfare is shown bymodifications todefenses
over the course

structed in separate episodes, or thickened with additional faces, and blocked doors and access

of pukara

lifetimes:

walls

con

routes. sturdy constructions, almost always built with a double face and rubble fill, totaling 1 to 2 m thick Colla However, pukaras were not invulnerable. and, where theiroriginal height is discernable, at At 37 of the 44 pukaras surveyed,we could not m least 1.5 high.At the largestpukaras theyare truly locate a present-day, or year-round spring, pond, cisternwithin thedefensive walls. While a limited massive, up to4 m thickand 5 m high. Some walls, especially rate on the more vulnerable sides, incorpo intent, for defensive

parapets?a

according toTopic and Topic (1987)?and occa sionally they includewatch posts or platforms.The presence of parapets points to the importance of

giveaway

supply ofwater could have been stored in largejars, it seems clear that most pukaras did not have the water supply necessary towithstand a prolonged siege. Colla populations must have relied instead on fortificationsto impede and discourage assaults

348LATIN AMERICAN

ANTIQUITY

[Vol.

19, No. 4, 2008

military and logistical capacity of surroundingsoci

eties.

or raids, and to delay attackerswhile allies could arrive.Prolonged sieges were probably beyond the

point of thehill, and may have formed an impor tantspatial and social focus for thecommunity and itsceremonies. The multifunctionality of these set
tlements demonstrates

While
were also

these siteswere clearly defensive, they

complete settlements where communi

ties lived and engaged in a multitude of activities. Nearly 90 percent (39 of 44) are associated with habitation areas, usually located inside thewalls. range from unoccupied refuges and modest sites Circular house foundations, 3-4 m indiameter, are with less thanhalf a hectare of surfacematerial to with up to 18 hectares of artifact usually visible on the surface.1 Surface artifacts large settlements from domestic occupation (ceramics, spindle scatter,or 300 to 600 houses. Judgingby theden some whorls, flakes, and grinding stones) are plentiful, sity of surface and excavation materials, and 10 test-pits foundevi placed inhouse structures dence of domestic use (including informalhearths
or ash lenses, use surfaces, faunal bone fragments, pukaras were

plete communities, not just garrisons of soldiers. Although Colla pukaras share aspects of gen eral design and the types of features they encom pass, there is also a great deal of variation. Pukaras

that pukaras

housed

com

and ceramic and lithicartifacts including spindle whorls). However, floorsare not highly compacted,
and

of timewhile otherswere only brieflyoccupied, if at all. Pukaras also vary a great deal in the strength of theirfortification walls and the accessibility of theirnatural landforms.Finally, the ratio between the totalvolume of defensive wall construction and the habitation area size (or number of structures)

occupied

over

a substantial

amount

carbon dates at some pukaras, the density of arti modifications topukara defenses, and use facts, the wear of grinding boulders, steps, etc., indicate that some pukaras were indeed occupied and used for a considerable length of time, but this does not appear tobe truefor individualhouses. Houses may
been re-leveled, have razed, terraces shored up and periodically and new houses instead of constructed,

floor identified in a single house. This preliminary evidence may suggest thathouses were not used intensivelyover long time periods. The spread of

in only

two cases

was

more

than one

potential

isquite variable, suggesting that while most pukaras could have been built by theirresidentpopulations, wall by wall, over a few dry seasons, othersmust have required a largerpool of labor. Keeping thisvariability inmind, we can tenta tively envision themultiple ways pukaras were used militarily. The largestwere the major politi cal centers in the region, protecting large popula
tions, small were livestock, communities temporary and stores behind massive walls, Some set

a single house being used repeatedly by multiple The internal layout of pukaras is variable and determined to a great degree by the topography of the hill. Most pukaras appear to be largely unplanned aggregations of house structuresfound At some larger pukaras, walled alleys meander throughthe terracesor compounds, imposingdirec
in groups, several to a line or a residential terrace. generations.

while the smallest shielded outlying families or

from refuges, small-scale and some, raids. permanent

tlements. Many may have functioned as retreatsfor additional vulnerable populations living below, who would have contributed to the construction
effort at some

made excellent sentryposts of the landscape, they for detecting enemy advances ahead of time, sig ally monitoring nearby activity. They were strategically placed to control surrounding lands, making itdifficult to encroach upon that territory without defeating thepukara. They formed visible and signals of a group's strength, were surelydeter rents for attack, yet itwould probably be inaccu rate to envision them engaged in a hostile but bloodless "cold war" that never included actual violence. Across cultures, decentralized societies that fortify their settlements engage inwar fre quently (Solometo 2006), and theongoing concern
naling to other pukara communities,2 and gener

pukaras.

With

commanding

views

and dividing the site tionon the flow of foot traffic into sectors that may have helped to define sepa rate social groups. Pukaras also include tombs as well as small storage structures, possible livestock
enclosures, nonresidential tures, when and sometimes, architecture. present, are other Small scattered special-purpose struc storage throughout the

residential areas, suggesting that each family or residentialgroupmanaged itsown surplus.By con
trast, tombs are

tions of the site, sometimes occupying thehighest

usually

segregated

in separate

sec

Arkush]

WAR, CHRONOLOGY,

AND CAUSALITY

INTHE TITICACA

BASIN

349

with pukara defense, manifested inwall augmen blocked access routes, and stockpiled sling tation,
stones, The speaks to a real and of pukaras sustained with war fear of attack. was strong association

(L6, Figure 5), thereare fivemajor and twominor five different clusters of storage structures, corrals, and six discrete clusters of tombs. Indeed, most
pukaras have two or more tomb clusters. It seems

nerability, by phrases for capturing defeated pukaras and destroying them (in fact, Bertonio gives four separate ways of saying "to destroy a
pukara").3

enough to leave traces inLudovico Bertonio's early colonial Aymara dictionary of 1612: their role as refuges is indicated by terms for fleeing and tak ing shelter in a pukara, and theiroccasional vul

likely thatdistinct social units inhabited pukaras simultaneously,keeping theiridentities separate in

through the maintenance of separate cemeter

ies. Farmland was also divided into segments at many pukaras, by verticalwalls runningdown the hillside thatprobably date to the same era (thever

part

Finally, it is notable simply how common pukaras are in this region (Figure 2). The overall distributionof pukaras demonstrates thatthe threat
of attack was pervasive. and ceramic Patterns styles of buffer examined zones, else fort clusters,

ticalwalls abut defensivewalls and are closely sim ilar in construction).Altiplano communities today use vertical walls partly to divide farmland into slices fordifferent social segments (Erickson 2000), so they have held the same purpose in thepast. may
In sum, there is evidence at this time for the per

where (Arkush 2005,2009) suggest that theColla riven into several politically region was
autonomous pukaras, each several hosting subregions a scenario consistent with relatively fre

quent local wars. The same appears true for the Lupacas (Frye 1997; Frye and de la Vega 2005; Stanish 2003). Yet thecontrastbetween thisarchae ological picture of fragmentation or very loose coalition and the ethnohistories of powerful war lords controlling large territoriesis perplexing. Following Stanish (2003:291-292), one possi bility is to envision Titicaca Basin Late Intermedi ate Period populations as segmentary societies, or used loosely here to refer togroup identity affil
along lines of perceived genealogical relatedness

sistence of smaller local identities, both within pukara sites (using distinct residential areas, ceme and teries,storage structures, fieldwalls) andwithin clusters of sites (using differentpukaras). larger This patternof decentralization suggests the main tenance of a degree of local autonomy even within largercooperative, and probably hierarchical, sys
tems of pukaras.

Pukara

Chronology

A pukara chronology can be drawn primarily from

carbon dates and

iation that is nested into larger and larger units

over time.Radiocarbon the course of fortification dates were obtained from a totalof 15 pukaras: at 10, testpits yielded samples of burnedwood from
occupation or midden contexts, and at eight pukaras

secondarily

from ceramics

to trace

(Evans-Pritchard 1940). In segmentary societies, (including threewhere test pits were also exca internaltensionsand conflictsroutinelycause larger vated), samples of dry grass (Stipa ichu)were col lected from interior mortar in defensive walls that groups to fission intoequivalent units or segments, but these segmentscan band together again intofor was exposed where wall sections or doors had midable defensive confederationswhen threatened fallen.These grass samples give dates specifically Andean kinship forconstructionor rebuildingepisodes of thedefen (Sahlins 1961).While traditional is flexible and based on bilateral descent, in con sive walls, and also avoid the potential curation
trast to the unilineal descent of "classic" segmen

tarysocieties,Andean corporate descent groups or

ayllus nevertheless are nested in

systems (Albarracin-Jordan 1996; Isko 1992; Piatt 1986,1987). These systemswere probably inplace by the end of theLate Intermediate Period ifnot before. Several traces of segmentaryorganization are detectable not only from thebroader landscape of pukaras but from internal subdivisions within
pukara sites. For instance, at the site of Apu Pucara

segmentary-like

problems ofwood charcoal in the largely tree-less environment of the altiplano.4 The resultingdates are shown inTable 2.

was some lim First, itshould be noted thatthere ited use of pukara hilltops in the Middle to Late Formative period. One pukara in thenorthernsur vey area (AS1, Calvario de Asillo) features con

siderable Middle and Late Formative surface ceramics and aMiddle Formativemidden that gives a date between about 800 and 540 cal B.C. (la).

350 Table 2. Radiocarbon Phase pre-LIP AA12871 AA12872 AA54218 LIP I AA54233 AA54220 AA54248 AA54237 AA54254 AA54246 LIPIIa AA54247 AA12875 AA12876 AA54226 AA54245 AA12874 AA56163 AA56162 AA54249 AA54255 AA54252 AA54234 AA54231 AA54242 AA54243 AA54241 AA54229 AA54228 AA12873 AA54244 AA54232 LIP lib AA54221 AA54223 AA54222 AA54217 AA54219 AA54225 AA54235 AA54236 AA54253 AA12877 AA54227 S4 CA2 CA2 AS5 AS5 AS5 AS1 PKP7 PKP7 AZ1 AZ1 AZ1 J10 Jll Jll AS1 AS3 N2 N2 N2 N2 N2 S4 S4 N2 L2 L6 L6 Dates

LATIN AMERICAN

ANTIQUITY with OxCal

[Vol. 19, No. 4, 2008 v3.8. Cal. A.D. 902-376 795-521 437-890 1017-1163 1028-1300 1030-1219 1071-1280 1068-1286 1185-1300 1277-1393 1042-1386 1158-1408 1243-1388 1279-1394 1296-1434 1277-1391 1281-1392 1282-1400 1298-1417 1284-1398 1297-1403 1284-1395 1300-1410 1286-1397 1293-1400 1285-1403 1303-1428 1280-1615 1296-1402 1304-1446 1314-1451 1324-1443 1330-1478 1401-1640 1328-1451 1332-1475 1328-1446 1331-1455 1409-1626 1327-1665

for Pukaras of theColla Area. Calibrated Age BP charcoal charcoal dry grass charcoal dry grass charcoal charcoal charcoal charcoal charcoal charcoal charcoal dry grass charcoal charcoal charcoal charcoal charcoal charcoal charcoal charcoal charcoal charcoal charcoal charcoal dry grass dry grass charcoal charcoal charcoal dry grass dry grass dry grass dry grass dry grass dry grass charcoal charcoal charcoal charcoal 2510 ?125 2520 ? 30 1370 ?100 955 ? 35 813 ? 81 896 ? 40 830 ? 35 815 ?44 766 ? 46 679 ? 34 780 ? 65 725 ? 80 709 ? 34 671?34 580 ? 60 680 ? 30 670 ? 30 652 ? 43 591 ?44 651?37 615 ?35 656 ? 34 596 ? 34 650 ? 34 631 ?34 639 ? 47 565 ? 34 545 ? 85 623 ? 34 530 ? 47 516 ?41 525 ? 34 487 ? 40 427? 70 502 ? 35 486 ? 38 510 ?34 491 ? 34 435 ? 42 405 ? 90

Lab No. Site_Context_Material AS1 midden, top AS1 midden, base AS3 AZ4 AZ4 P29 P29) P29 defensive wall mortar (bad date?)a poss. house 2 floor defensive wall mortar terrace fill midden terrace fill L2 midden fill below house house floor and below L6 midden defensive wall mortar terrace fill L2 midden fill below hearth hearth inside house S4 midden terrace fill house floor house floor house floor poss. house floor fill below house fill below house defensive wall mortar defensive wall mortar house 1 floor fill below house 2 house 2 floor defensive wall mortar defensive wall mortar defensive wall mortar defensive wall mortar defensive wall mortar defensive wall mortar exterior use surface house floor house floor poss. floor defensive wall mortar

513C -23.9 -23.2 -25.9 (-25) -23.2 -22.7 -24.1 -24.2 -23.9 -23.9 -24.5 -23.9 -24.4 -22.6 -23 -22.8 -24.9 -27.3 -24 -24.4 -24.9 -24.8 -23.7 -23.7 -24.6 -22.6 -24.3 -23.2 -23.6 -23.9 -22.2 -24.6 -17.3 -25.2 -24.2 -23.6 -24.2 -22.7 -23.8 -23.5

Cal. A.D. 794-114 788-545 564-775

(la) BC BC

(2a) BC BC

1024-1122 1072-1284 1042-1209 1188-1259 1191-1274 1223-1284 1282-1386 1190-1292 1218-1388 1274-1379 1284-1387 1305-1411 1282-1384 1285-1387 1294-1388 1306-1403 1295-1388 1303-1394 1292-1388 1306-1400 1296-1388 1300-1390 1298-1391 1323-1417 1305-1438 1301-1394 1330-1437 1335-1438 1401-1434 1408-1444 1418-1622 1409-1438 1411-1443 1407-1436 1412-1440 1428-1484 1432-1630

-11 1422-1467 1408-1487 PKP2 455 ? 34 dry grass -24 AZ4 hearth below house 1 383 ? 59 1444-1626 charcoal 1436-1640 _AA54251 from the outer wall at the same site, which was in line with aCompare this date, for themiddle wall, with sample AA54219 other pukara wall dates. The two walls are similar in condition and appearance.

Four additional sites in the southernbasin also have considerable Late Formative ceramics. There is no

mediate Period, before approximately A.D. 1275, when few pukaras were built or used, and a phase thatdefensive walls were constructed in of greatly expanded pukara use in the late Late evidence Intermediate Period and possibly beyond. (This the early period, but it may be significant thathill as were second phase can also be split into two subphases sometimes However, tops occupied. use dates to the Late at around cal A.D. 1400 based on carbon date dis most Inter expected, pukara
tributions, dates allow us to separate Late Inter between two phases: the early but the there are no dramatic used changes in these The radiocarbon use into subphases.) Pukaras

mediate Period.
pukara

phases are shown inTable 3, along with their site

>

CW 3;

3JOI 3JZo rOQ > Z a o > c > O

H z HXm HO

> > w

en CO

House

and

Min.

ascent

puks Visible

Visible

puks

Max wall

Max

Number wall wall Total

average 1.6 time, 2.0 4037 2.5 3.3 1000 35 2.0 4.0

2 6.2 within km 2.4

within 4132 4.2 3.1 1846

40

2.7 7.4 average_5,9_2/7_3,4_2086_2$_4146_4]_3/7_

Table Summarized Information 3. Radiocarbon Pukaras with Dates. on

(m)_length (m) Altitude_(minutes)_10 height Site_area Phase (ha) thickness of walls km_20 structs km

Llongo(S4) 4.8 2200 3.4 1.8 77 4010 35 0 Cerro Pucarani (N2) 2100 4.6 188 5.65 2 4300 52 1 Cerro Mugra(Jll)

800 3.4 2.1

440608 42 Cerro Pucarani (N2) 2100 4.62 5.65 18 2 4300 52 1

3.0

IMuyu Pucara (AZ4) 3.5 7 143 339 2.7 1500 3 4210 4

Ha2.8 7 Apu4 Pucara (L6) 232 1.75 2.3 2500 7 4220 44

K'askawi (AS5) 3.1 5 158 1.9 1,5 1900 3 4040 33 3

11 8 35 4120 3 5950 2.9 2.3 324 8.2 (AS3) K'ajro Cerro

Muyu Pucara (AZ4) 3.5 7 143 339 2.7 1500 3 4210 4

(PKP7) 6 3 49 4260 Mallacasi 3 Cerro 2.8 2.1 1000 4.7 30 4 44 1 4200 (CA2) Sinuchache Cerro 360 2.5 555 7.6 1.8 4 38 900 3.5 4130 CerroMonos(JlO) 6 2.5 2.2 lib 1 (PKP2) 3 41 Chico 4080 Kojra 200 4.5 .7 CalvariodeAsillo(ASl) CalvariodeAsillo(ASl) 1.5 Pucara 5.6 Ch'unchu 1.9 3150 (AZ1) 4 13 9.8 4120 44090 37 5 1.9 1.5 5.6 3150 4 13 4120 37 5 4.2 2600 5 14 39 6

31 1 (P29)2.4 2 Yanaperqa 36 Pichuni 1 (L2) 3950.9 3 700 1.9 Cerro 2.3 3 .9 800 Toclomaro 4 .5 3950

3 31 (L2) 1 3950 Pichuni 800 2.4 .9 1.9 Yanaperqa

352LATIN AMERICAN I' ' l-i.--T"?,?j?-"\ \

ANTIQUITY

[Vol. ': ' :',7K-^^p j

19, No. 4, 2008

rlA-vl if.
!

'

-A : V \ \ \ ! ?^*>*#^M ***MKil i / \ \ \ VV A \ \ \
I >?s\ ?c*i#iiiiiPi8lli!

: \-\

| ]

i'

lf?

^%

\ \

/^d*t/1m*?}0**m:

j I^^^M

V-

? II |

- te9*M^ :'/^/0

Im^Kk

co/

' \?/
A 75 60

/
100 m

//
A |T
(P29).

MMC=B?M " 0 25

Contour lines approximate.

Figure 3. Cerro Toclomaro

acces size,wall height and thickness,and landform All indices of defensibility increase over sibility.5 time,but because of the very small sample size of pukaras fromphase I, only the change in site size Here I fromphase I to II is statistically significant.
describe the contrasts between

pukara on a low but steep hill just northeastofmod ernPuno (Figure 3). Relatively low defensive walls supplement a rocky ridgetopwhere 16 chullpas are perched. No circular house foundations are visible
scatter on the western encountered side of the site, and a midden. This test-pit pukara,

at the site,but there is about .5ha of dense artifact

excavations

give a few examples of individual pukaras to illus

trate.

the two phases,

and

which could only have housed a small population, was probably abandoned by theend of thisphase,
since there is no trace from the surface or excava

Phase

I: A.D.

1000 to 1275

In thefirstphase of theLate Intermediate Period,

pukara use was minimal.

the sample of 15were used at this time, and none of themare large.Cerro Toclomaro (P29) isa small

Only

three pukaras

from

tions of Sillustani pottery,a ceramic type that is common inthisarea in later Late Inter phases of the mediate Period. Second was Muyu Pukara (AZ4), a site on a ridgetop in a chain of high hills north
east of Azangaro, an unusual pukara in several

Arkush]

WAR, CHRONOLOGY,

AND CAUSALITY

INTHE TITICACA

BASIN

353

?%

' ^?'"

;'*?--^?-r'

<

; - *^#$?^

-"

' A

'' :^'i:J^^^^'&-'

'

'/

^^^'.v.:W./..\\*.:.
' '*"*" -"". T . . ' -:';. -I...?..... i. ,ii ii . "" ,??? ?.,.?, ?.,.. "'.."-? ii ?... ..iii?II.I..MI. t'-i ii.i,?iii,i., / ' ,..., ,, *' M?pp?d%vW>^caMar5AMr?frfi?Ki?o17$27.pn)^ct455-97-A Pucara (AZ4).

'? ,J": '..' '^'.w.'JjWi

6' " 50"10^ W^^^MOm "' " "" .?'" .

Figure 4. Muyu

respects (Figure 4). It has over a hundred house

foundations in an area

built defensive walls pierced by numerous doors.

However, surface, artifacts and are on the nonexistent practically test excavations inside two houses and

of 3.5 ha, protected

by well

tinuouslyor periodically occupied from some time after A.D. 1200 intophase II of theLate Interme diate Period. Thus, in phase I there is no evidence of large, defensive, permanentlyoccupied pukaras. strongly A few brieflyoccupied or less defensible pukaras
were out used, and the survey area the fact that they are spread that a degree suggests through of con

inone cleaned house profile found no distinct use

surfaces

was used sitewas used only briefly. Nevertheless, it more than once: dates from a defensive wall and an ash lens inone house fall inphase I of theLate Intermediate Period, but the site also had at least one lateroccupation very late in theLate Interme diate Period or Late Horizon. Thus, it is not known how much of the site dates to the early phase. Finally, Pichuni Yanaperqa (L2) is a small pukara with 1.9 ha of artifact scatteron a very low, easily

and very

few artifacts,

suggesting

that the

flictwas present regionally. Phase II: A.D.

1275 to 1450+

The fourteenthcenturywitnessed an explosion in nine out of the 15 sampled pukaras were in use, with an additional five falling afterA.D. 1400. These sites include a wide range of sizes and defen
sive strengths; four examples illustrate their vari pukara use and defensive wall construction. At least

accessible hill southeast of modern Lampa. The walls at the top of the site are the best defined, reaching up to 2.4 m inheight.Here, entrances are atively low priority at this site, although at some point the higher entrance on the west side was was con blocked. Dates from this site indicate it
aligned, an indication that defense was a compar

ability. CerroMallacasi (PKP7) is a small pukara north west of the town of Pucara with only about 30
house structures

2 ha (see also Kidder 1943). Nevertheless, the site

dispersed

on

low

terraces

across

354LATIN AMERICAN

ANTIQUITY ..

[Vol. . . i .-.".-mi1-''^ ??''i','^,-0':'^^

19, No. 4, 2008

y'

ancient VWQCti '*'

. *"*?

-^^r-.:'-

.:'$

^''

|AASAAA;' '^fi-'^-^j^J^

>tombs-t '.unmapped

Nfi^ ^?,

^^ \ A^

V^ /^s^.'

^^^^V^

>*"

r-

:-^

'*?"^ ^-**->s--f

&."

\ _
- TOO O 200 300 5. Apu Pucara * ^,, 400m Figure (L6).

-'-:

/-' ,.4
!

-.. :

was placed on a highly defensible, steep hill pro tected by sheer cliffs on three sides. A massive wall bars thefourthside.The wall has caches triple
of

high and furnishedwith parapets. The gate opens into a walled path that leads to the top of the hill, passing between terraces with about 150 house
foundations and 75 smaller

ways, one of which was later blocked from the inside.Boulders carved with petroglyphs are inter spersed among thehouses, and tombs populate the bedrock ridge above. With walls and cliffs that
enclose

slingstones,

parapets,

and

small,

baffled

door

west is a wide, flat saddle, with side the walls to the

five corrals again, posed that and six clusters of tombs, this defensive social community segments. suggesting, was com

storage

structures.

Out

of smaller

The great disproportion between the size of the defensive walls and the handful of houses at this helped build the walls, unless the construction dragged on for decades.6 The pukara probably offered a refuge to nearby unfortified populations who participated in its construction and mainte
nance. pukara suggests that non-site residents must have

swale ideal for pasture, this pukara was better most to withstand a prolonged siege. equipped than

a year-round

spring

and

large,

grassy

Cerro K'ajro (AS3), situated on a long, rocky ridgetop east of the townofAsillo, is a good exam ple of a largepukara (Figure 6).While the ridge is easier toascend than most pukara landforms in this thesite isprotectedby threetall walls ingood phase, condition, reaching 2.9 m in height and 2.3 m in thickness on the east end of the site where the approach is easiest. This site displays clear evi
dence of defensive

aligned doorways, and a cache of rivercobbles for

use as

design:

parapets,

small,

non

Apu Pucara (L6) is a pukara ofmedium size on a high, steep hillwest of Lampa (Figure 5). To the
east

walls are 323 circular house foundations inan area

of about 8.5 ha covered

slingstones

by

the western

walls.

Inside

the

line theother sides. On the west side of the site, the is flankedby inset walls about 2.1 m principal gate

is a steep,

rocky

drop;

seven

concentric

walls

Houses on the southern slope are grouped in lines

on shallow terraces; those near

by dense

surface

ceramics.

associated with several compounds defined by a

the ridgetop

are

Arkush]

WAR, CHRONOLOGY,

AND CAUSALITY

INTHE TITICACA

BASIN

355

--^

100

.,.;.T^,

aoo fa

air 455-97-A MappedwiththeaidofSAN photo 15600. project Figure 6. Cerro C'acjro (AS3).

single course of boulders, which may have been

livestock pens.

at thevery end of theLate IntermediatePeriod after

cal. A.D. indicate To 1400, when the site was summarize, dates from two test units also occupied. this second phase of the Late

Pucarani (N2), the largest and most strongly defended pukara in the survey, is located on and around the sloping peak of a steep massif, pro
tected

monumental walls defend the (Figure 7). Two long,

western

by

cliffs

on

the north,

east,

and

southeast

IntermediatePeriod saw the height of pukara use in the northernbasin. While some small pukaras such as Pichuni Yanaperqa (L2) and Mallacasi
(PKP7) medium were used, others, such as Toclomaro

theeastern cliff; thesewalls average 3.5 m high and 3.7 m thick,and reach 5.6 m high and 4.6 m thick
at maximum. row doors, They pierced some of which have are by numerous nar small guardrooms

approach,

and

another

wall

bars

access

to

(P29), had been abandoned, giving way tomany

and

built into thewalls on one or both sides. A parapet and inner walkway are well defined on the outer
western door wall, and the walkway creating is cut wherever elevated stations a on passes through,

These largersites often have very high, thick walls, and are found on more defensible landforms, indi
cating sites, an escalation separate of the threat of attack. At compounds some residential or other

large

sites with

hundreds

of houses.

indices suggest thepossible nucleation of smaller

social

both sides fromwhich tomonitor anyone passing

in or out. While probable surface over only some storage structures habitation house are foundations and the distinguishable,

cates that pukara inhabitants probably formed

defensive coalitions,

segments.

In addition,

pukara

clustering

indi

artifact

scatter extends

shallow

and below the easternwall. If itfollowed theden sitydistributionof otherpukara sites,Pucarani may
have included a thousand or more house structures.

18 ha approximately terraces on the sloping mesa

the level of the site and thereby allowing some smalleror lessdefensive pukaras to survive (Arkush 2005, 2009). Carbon date distributions allow us to see that
pukara

maintaining

social

units above

The spread of dates from this site indicates an

extended

thevery end of theLate IntermediatePeriod, after

1400. As a group, these late pukaras are

occupation

and construction

continued

into

cal A.D.

was built just at thebeginning ofwidespread pukara use, in the final decades of the thirteenth century. wall was builtor rebuilt The southern portionof this

period

of use. The

outermost

western

wall

not clearly distinct in typeor number frompukaras used in the fourteenth century.What they do demonstrate is a sustained level of the threat vio of
lent conflict over a certain time frame. By the end

356LATIN AMERICAN

ANTIQUITY

"~~

[Vol.

19, No. 4, 2008

- c<*
^

^ftk-.

; ?--IT

;<4flm,*?,*'?W#.1.

?0

'' ?' 200 100'

N1-.

?,

t^

-"

* / X

300~

^?^-.^^^ 400m
Figure 7. Pucarani

^
ICT*

+* doofwithgo&uoorns toboth sides (N2).

or extended nature of pukara occu Mallacasi (PKP7), Kojra (PKP3), Cerro Mugra episodic and construction is suggested not just by the (J11), and Cerro Pucara (V3). These later improve pation time frame of the carbon dates, but by a number of ments to fortificationssuggest not only long-term or episodic use, but thattheperceived level of threat at at least two occupations other clues. There were remained high over time.Not every site remained Cerro Sinucache (CA2) and Chila (V2), shown by a reliable stronghold to be reused later,however. house and wall superpositions. Threshholds and The

of theLate IntermediatePeriod, in thefirsthalf of of thefifteenth century,the threat attack on pukara had not lessened. populations

theaddition of a new layer to theexterior (dressed) face of a wall to thicken it. Blocked doors indefen sive walls are visible at Pichuni Yanaperqa (L2),

sage of people. Multiple wall-building episodes at Pucarani (N2) have already been noted; atCaritani (LI) and Pucara Orqo (AS4), theyare indicated by

stairways

at numerous

sites were

worn

with

the pas

As

has

been

noted,

at

least

one

site

P29) was abandoned by the second phase of the Late IntermediatePeriod, and walls at several sites While the may have been intentionallydestroyed.7

(Toclomaro,

Arkush]

THETITICACA ANDCAUSALITYIN BASIN WAR,CHRONOLOGY,

357

local histories of individual pukaras cannot be wholly untangled at thispoint, it is clear that pukara
use was violent a process, attack waxed not a brief and waned event. over The threat of of the course

several generations. Through victories and defeats, fortsendured to be used again by thosewho could
control them.

crag above (PKP5). There are unusually late dates forwall construction events at both Calvario de Asillo (AS1) and Cerro Kojra Chico (PKP2) at Llallagua, and possibly these dates correspond to theColla rebellion.
Inca architecture is also

tion and governance of thisbellicose region (Fig ure 8). These architectural complexes are located, again, on the rebel strongholds of Llallagua and The very late range of dates such as the one from Asillo. There was a considerable Inca investment inarchitectureat theLlallagua hill range,with four Muyu Pukara (AZ4) brings up the issue of pukara use at or after the Inca conquest. use Late Horizon Inca installations,one on or close to each fortified appears tohave been the exception rather than the peak. The largest is a complex of niched rooms rule: distinctive Inca-period ceramics are plentiful around a courtyard on theLlallagua peak proper, in theTiticaca Basin generally, so the paucity of first described by Rowe (1942), who proposed that Swords and Ploughshares: Pukaras in the Late Horizon them at Colla pukaras suggests that most pukaras were abandoned in the Late Horizon. Large pukaras are often found near Late Horizon centers like
the structure was

attesting to a presence thatwas probably official and that illuminates Inca strategies of incorpora

present

at a few pukaras,

Asillo, Lampa, Nicasio, andVilque, and probably supplied theiroriginal population.

However, on some late, attenuated seven

niched rooms with a patio and enclosed bedrock platform are found on the secondary peak ofCerro Kojra, a small group ofmasonry-retained platforms overlooks theapproach toCerro Kojra Chico from Llallagua's peak, and finally, there is a poorly pre served compound of rectangular rooms around a patio on the lower, lightly fortifiedpeak of Qui
jache. masonry Niches, rectangular these layouts, as and shaped identify structures Inca. At Cal

an Inca garrison.

In addition,

two

burial is indicated by Inca-period chullpas or

ceramics at least out of the 44

occupation

or

For instance, small but significantamounts ofLate

Horizon ceramics are found at the sites of Cerro

pukaras.

(N2) and Mallacasi (PKP7) described above. It is intriguing to speculate whether such material is related to the defensive use of pukaras at the time of the Inca invasion, or in the rebellion that,according to the chronicles, took place after
in the Colla area at false news of the Inca's

Pucarani

wards

cance in the Late IntermediatePeriod, with several petroglyphs and a large circular enclosure. It is Llallagua, and includes a facade with double jambed niches on one side and simple niches on the other side. This set of structures is especially because the impression itgives is of a interesting
ceremonial larger and more complex than the structures on

vario de Asillo (AS 1), the Inca architecturalgroup is located on a saddle to theeast of thefortified peak, in an area that may have held ceremonial signifi

death. The chroniclers state thatthe rebellion took

some

1988:130 [1572:Ch. 50]). Most of these rebel pukaras can be identified,except for the case of Arapa. Calvario de Asillo (AS1) was almost cer tainly the rebel fortofAsillo, being the largestand closest pukara to the town, while "Llallagua" must have referred to at least one of the pukaras in the complex of four (PK1-4) on a massif south of named for therebel strongholdon the Pucara, itself

quell, and they name specific pukaras used (or reused) by the rebels in thisuprising: Pucara, Llal lagua,Asillo, andArapa (Betanzos 1996:144-146 1951:336 [1551-1557: Pt. 1, Ch. 34]; Cabello [1586: Bk. 3, Ch. 18];Murua 1986:89 [1613:Bk. Ch. 24]; Rowe 1985:213-214; Santa Cruz Pacha 1, cuti 1993:234-236 Sarmiento [1613:26-27];

time, perhaps

years,

for Topa

Inca's

army

to

Double-jambed niches and doors are extremely rare in theTiticaca Basin, with themost salient exception being the major temple complexes on the Islands of theSun and Moon. In this light,thefunc tion of the Inca structureson Llallagua could also
reconsidered. be

complex,

rather

than

garrison.

The most importantof the rebel pukaras was Pucara itself,and here the Incas also made a sig nificant building investment,not on the fortified peak, but in alterations to the Formative period ceremonial site at thebase of thepeak. Topa Inca incorporatedPucara andAsillo (and possibly Llal

358LATIN AMERICAN

ANTIQUITY iujc^. i

[Vol. UHfa9ii?W???nptee

19, No. 4, 2008

Hmun tadap<edfromRowe1942)

/N

JPMW

'H.|

I
#

lj_L^
Llaftaguanorth
I Ml LWIagua complex:

JL^hI LKnoJ

11

It

Uaaaguahffl

complex:

ljljj T Q

?n ir1

a;:;..:?; v;
ctKSf/SiM$Ctm^SMi

T /

I fublAM^cd platform

......

r"Aii~i
' complexes

IsBSHBpBBHHHiSHi

Figure

8. Inca architectural

on Colla

pukaras.

lagua) into a royal estate after vanquishing the Colla rebels (Rostworowski 1970:162), so Inca structuresand modifications may have been built
as an

tion thatwas interestingly multifaceted: military trativecontrol as part of a royal landholding, and the symbolic appropriation or neutralization of rebel forts through the building of ceremonial
structures. reconquest, the establishment of direct adminis

with Colla rebel thejuxtaposition of Inca structures

pukaras reveals an Inca response to the insurrec

improvement

to these

royal holdings.

In sum,

Arkush]

WAR, CHRONOLOGY,

AND CAUSALITY

INTHE TITICACA

BASIN

359

Discussion: Causes of War While

across

The Course and in the Titicaca Basin

the evidence indicates complex local histo

construction, region use, and modification, a dramatic esca the Colla there was

ries of pukara

lationof the threat war in the fourteenth of century. Fort use and wall construction continued apace of afterca.A.D. 1400,while thedefensive strength
forts remained the same or increased. Two fort dates

the late Late Intermediate Period. Consequently we must reconsider the explanations for it. Collapse

While more dates regions at about the same time. are needed from the central and northern Andes, the oft-mentionedwarfare of theLate Intermedi ate Period seems to be mainly a phenomenon of

In the upper Mantaro basin of the south-central Andes, defense was already a concern in the early Late Intermediate Period, but warfare apparently A.D. 1300, when Wanka popula intensifiedafter tions congregated inmuch larger,higher-altitude settlementsbehind thick walls (D'Altroy and Has torf2001; Earle et al. 1980 1987; Hastorf et al. 1989). Likewise, in the upper Moquegua valley, some Tumilaca populations of theearly Late Inter mediate Period built fortifiedsettlements, signal inga concernwith warfare, but nearly all sites from the succeeding Estuquina period, dating after cal A.D. 1200 (with the majority aftercal A.D. 1300), were fortified and defensibly located (Owen 1995; Stanish 1985, 1992). Warfare and fortification in

beyond theTiticaca Basin, for regional studies in Andes severalotherparts of the southernand central bear a telling resemblance to theColla sequence.

probably fall after cal A.D. 1450 and may corre spond to theColla rebellion. What caused the intensifiedwar of the four teenth century? Whatever itwas, it stretched far

Recent reevaluations of dates indicate that Tiwanaku's collapse occurred around A.D. 1000 Wari's at approximately the same (Owen 2005) and

time (Williams 2001). If theextensive fortification of the Titicaca Basin and thebroaderAndean high landswas a direct resultof thecollapse of the Mid dle Horizon states, it should have occurred in the firstfew generations afterwards.Clearly thiswas not the case in the northern Titicaca Basin. While the use of a few small and less-defensive pukaras early in theLate Intermediate Period indicates a most fortification occurred after degree of conflict,
a delay allowed of at least two centuries. We can consider

political collapse to have been a precondition that

endemic warfare to occur. However,

another explanation isneeded for the marked esca lation in warfare in the late Late Intermediate Period. Invasion
If not a result of collapse, were pukaras a response

to Inca aggression? Colla pukaras arose and became was the far southern Andes a late clearly phenom widespread significantly earlier than the Inca enon. Nielsen's (2002) dates from pukaras in the we adhere toRowe's (1945) conven expansion, if Altiplano de Lipez inBolivia's southernhighlands tional chronology,which places the Inca conquest correspond closely to the height of Colla pukara of the Titicaca Basin at aroundA.D. 1450.This tim use, with five out of six falling between cal A.D. ing is supportedby radiocarbon dates from the Inca 1300 and 1400. To the south,pukaras near Chile's shrines on Lake Titicaca's Islands of the Sun and Loa Valley and the Salar de Atacama date to the Moon, which fall within the late 1400s and early end of theLate IntermediatePeriod, aftercal A.D. 1500s with 1-sigma ranges after about cal A.D. 1300 (Uribe 2002). In northwest Argentina, settle 1420 (Bauer and Stanish 2001:251-255), as well ments moved tomore defensible and partly forti as by dates fromEstuquina-Inca and Inca sites in fied positions after A.D. 1200, reaching a height in theupperMoquegua Valley, also falling in thefif the fourteenthcentury (Nielsen 2001). Studies of teenth century (Stanish and Rice 1989:8; Stanish individual pukaras also usually yield dates in the dates for 2003:208). However, fourteenth-century latter half of theLate IntermediatePeriod, as at the strata with Inca ceramics fromCaquiaviri south of Titicaca Basin forts of Cutimbo (Frye and de la Lake Titicaca (Parssinen and Siirriainen 1997) sug Vega 2005:178) and Pucarpata (Parssinen 2005), gest conquest was preceded by contact and to the south of the study area. It is intriguingthat these developments occurred in such far-flung military forays took place as well.
exchange with Cuzco. Possibly, unsuccessful Inca

360LATIN AMERICAN However, pukaras are pervasive throughout the

ANTIQUITY

[Vol.

19, No. 4, 2008

Environmental
Environmental crisis and

pukara use spanned at least 200 years, and some siteswitnessed extended use. Thus, while some of the laterpukara dates may correspond to the Inca invasion or theColla rebellion, it is unlikely that most pukaras were built toguard against Inca incur
sions.

removed Titicaca Basin, even inpartsmuch farther from the Inca threat than the northern Colla area, such as the southern Lupaca zone. In addition,

Change
resource scarcity?

particularly drought?is an attractive explanation because itapplies well to the scale of theproblem. Precipitation is an extremely important factor in agricultural production in theTiticaca Basin: the terraced hillside and hill-base fields are primarily dependent on rainfall and runoffrather than irri Colonial observers in the altiplano noted a deep concern with crop shortages and famine in years
gation from permanent springs or rivers. Early

a late arrival to the Titicaca Basin (Albo 1987; 1987; Cerron-Palomino 2000; Bouysse-Cassagne 1980,1987; Torero 1987,1992; see Brow Espinoza man [1994] and Stanish [2003:222-223] for main overviews). Alfredo Torero (1987,1992) is the Aymara proponent of this idea, suggesting that displaced indigenous Uru and speaking migrants Pukina-speaking peoples who had been responsi ble for earlier cultures in the Basin, including Tiwanaku. While originally proposing that this

A more plausible explanation could be an Aymara invasionhypothesized by linguistsand eth nohistorians,who have argued thattherelativeuni was of Aymara language indicates thatit formity the

affectsaltiplano potato yields precipitation strongly (Garreaud and Aceituno 2001; Orlove et al. 2000:71).

of low rainfall (Cieza 1984:272 [1553]; Diez 1964:147, 163, 175 [1567]; Polo de Ondegardo 1990:61 [1571]), and recent research confirms that

Precipitation in theTiticaca Basin, like the rest of the southern Andes, depends mainly on north easterly winds from thewarm, humid Amazon Basin during the austral summer fromDecember March. It is highly variable from year to through year and on longer time scales. It is substantially lower during severe ENSO episodes (El Nino? Southern Oscillation) (Roche et al. 1992), itoscil lates independentlyon a intervalof about 13 years (Melice and Roucou
by ice cores

Lupaca-Colla rivalry that the chroniclers describe as a conflict between invading Aymara and native
Pukina speakers. (Most Titicaca Basin archaeolo

influx occurred at Tiwanaku's collapse, he later the revised the timingto ca.A.D. 1200, interpreting

1998), and longer periods of and high rainfall are docu alternating drought
and lake cores

gists do not subscribe toTorero's hypothesis, citing

continuous

mented

continuities [Albarracin-Jordan 1996; Browman 1994; Stanish 2003; Stanish et al. 1997]). Ifpukaras were the result of an Aymara migration into the region, themigration must have taken place in a

site occupation

and

long-term

stylistic

that they are too local, failing to address the very and defensive settlement large scale of fortification in the Andean highlands in theLate Intermediate Period. If anAymara migration is seen as only one
of a number

complex patchwork thatcaused pervasive conflict throughouttheTiticaca Basin, ratherthanresulting in a single, archaeologically visible frontier. The main weakness of both early Inca incursions and an Aymara migration as causes of warfare is

1997; Baker et al 2001; Thompson et al. 1985, 1986,1998). Evidence fromcores of the Quelccaya ice cap, just north of the Titicaca watershed (Thompson et al. 1985), supported by sediment cores fromLake Wifiaymarka, the smaller lake at the southeastern end ofLake Titicaca (Abbot et al. 1997), indicate an extended drought period during the Late Intermediate Period. Binford and col leagues (1997) estimate thatprecipitation during mod thisepisode was 10 to 15 percent lower than ern levels, and lake levels dropped by 12 to 17m.

(Abbot

et al.

problem, the migration hypothesismust be reduced

to an intermediary step from other explanations.

placements taking place at this time, the question is simplypushed back one step:why was thisa time of extensive migration? To satisfy the scale of the

of population

movements

and

dis

in timebetween A.D.I 030 to 1280,with thereturn of shallow water by about A.D. 1350, and deeper water about a century later (Abbott et al. 1997). Oxygen isotopes in ostracods record the balance

would have posed an unprecedented Such a drought of challenge to the farmer-herders the altiplano. How well does this drought correlate with pukara use? According to the lake cores, a hiatus indicating a low lake stand occurred at some point

Arkush] .i . . ;

WAR, CHRONOLOGY,

AND CAUSALITY

INTHE TITICACA

BASIN

361

-.[, w.i.i.-iiii.ii.i.iii.i.ii..., j j ^..[.'.i^ .." ;w.:Mi^^q:|l^['^wu.-P;n.^i,ii.i.i.i... ?) n-i . ii.i.l,.-r?.?...-i.i.ii.ii.?i?ii,iiiim

1100 1000

1200

148& 1300

H00

fiatiAD
core 1 ice accumulation (above) and 8,80 (below) forA.D. 1000-1500. Figure 9. Quelccaya Trend lines show moving average. Dotted lines indicate averages for entire core sequence. Data from Thompson (1992).

of evaporation and precipitation, and likewise indi cate dry conditions after thehiatus until aboutA.D. 1400 (Binford et al. 1997). Unfortunately, theonset

of the low lake stand cannot be dated directly, because the lake core sedimentswere eroded while exposed in thedrought. The Quelccaya glacier cores, with theirannual
a much more precise chronol resolution, provide

ogy than thecarbon dates thatform thebasis of the lake core chronology, but their interpretation is
more

Basin beginning ofmost pukara use in thenorthern (Figure 10). However, pukara use would have mainly occurred after the drought eased, in the fourteenthand fifteenth centuries.This correlation should be considered tentative,because thephys icalmeasurement of ice layer thickness is prone to error (Calaway 2005; Melice and Roucou 1998).8 The Quelccaya cores also recordoxygen isotope ratios (8I80), which reflecttemperatureoverAma
and ture as well the tropical Pacific as precipitation sea over surface tempera the Amazonian

1985:973). Itwould have coincided with the very

ologists drawing on the Quelccaya data (e.g.Mose ley 2002; Ortloff and Kolata 1993; Shimada et al. Williams 2002) use ice layer thickness as an 1991; index of precipitation, and draw primarily on the
accumulation

problematic.

Thompson

and

the many

archae

zonia

sure,core 1 indicates thattheperiod between about A.D. 1240 and 1310was a drought of exceptional length and severity (Figure 9; Thompson et al.

sequence

from core

1. By

this mea

basin. The degree towhich one or theother factor affects isotope ratios on shortand long time scales is imperfectlyunderstood (Baker et al 2001; Has tenrathet al. 2004; Hoffman et al. 2003; Thomp son et al 2000,2003; Vuille et al 2003); temperature is the dominant factor on very long (e.g. millen nial) time scales, while variability on decadal time

362

LATIN AMERICAN

ANTIQUITY

[Vol. 19, No. 4, 2008

Colla pukaras: radiocarbon dates

Drought periods, from Quelccaya iceaccumulation iAJfcJ?D^B_ A aflfl^L_

AZ4 house 955?35BP) (AA54233 AZ4wall(AA54220813?81BP)_.

-^_j_

L2(AA54246766?46BP) _|_j __._XJ^L_ ?_~ L2(AA54245671?34BP) ---rU^

L2(AA12874580?60BP) ._

_.

_LJ-IHtfpfe^?_I

2 S4unit(combined dates) S4unit(combined 1 dates)

CA2 dates) (combined

IJj^B-J^-_ ?_ 4-|-|^HAA|

WXSsMi

?i|11 1i11 I^Bfc J^

PWSmMM

AS5(combined dates) _|^H^?^,_

AS1 house(AA54241 ..; 631?34BP) "1 %j3f^M01 ?J1 j |WKUKfk JJUS:t AS1wall(AA54217427?70BP) ||| |||| J^K_J_i^H^^^^_
PKP7 wait(combined dates)

AZ1 (combined dates)

-4^_

fill J10 (AA54221 BP) _. wall 516?41

. |^B|JB|A_ ^F=^ JH_

J11 (combined wall dates)

AS3wall(AA54219 502?35BP) _

wmm^Smii I) ISIJ1S^Htill "l^lM|||pS^l^^/^

H~^ ~j^"_

N2wall(AA54226709?34BP) ':;-^y^r-*

N2wall(AA54225486?38BP) |j N2unit2(combineddates)
900

^| |J|g | ^K-S

ftr

'

N2unit1(AA54235510?34BP) ||Ij B I^^^^Jl lj }S

lj ||||
IH_ PKP2wall(AA54227455?34BP) _ -db^-_

| ?jj| Mpl f^B?

^ 3Jj^S^5^^^^'

11 ^j^;:;...,:r"::BM&

H9||^ffr^-'fM^^^^^y:;

1000 1100 1200 1300 1400 1500 1600 Date AD

between pukara dates (black shapes) and peak droughts (vertical gray bars), as indicated by Figure 10. Comparison core 1 (Thompson ice accumulation, 1992). Droughts are defined as years in which the 10 yr running mean Quelccaya - lo for whole core < avg sequence.

Arkush]

WAR, CHRONOLOGY,

AND CAUSALITY

INTHE TITICACA

BASIN

363

2003). Oxygen isotope values from the Quelccaya core 1are shown inFigure 9; higher (less negative) values should correspond to drier conditions on

scales appears more affected by precipitation. For instance, for the relatively short timeperiod since 1915, levels ofLake Titicaca correlatemore closely to changes in 8180 in theQuelccaya cores than to ice layer thickness (Baker et al. 2001; Hastenrath et al. 2004; Melice and Roucou 1998;Rigsby et al.

arsenal ofAndean peoples indealing with adverse conditions. These aggressive strategieswere not had a powerfulmomentum of determined, but they theirown: if some groups chose them frequently enough, otherswould have been compelled to fol
low, or face unrelenting depredations.

short time scales. By this measure, dry and warm most of theLate conditions prevailed throughout Intermediate Period, includingdroughts fromabout 1305 to 1380 during most of the firstphase of
occupation, pukara suggesting use, resource a more stress, direct and link resource

Warmongering, Fortified Landscapes, and Durable Antagonisms Another kind of explanation is needed for theper Titicaca Basin petuation ofwarfare in thenorthern a time of ameliorated cli in the fifteenth century,
mate:

pukara between

unpredictability (Figure 11). One final lineof evidence is the Miraflores flood, a severeENSO event in theearly to mid-fourteenth evident in flood deposits from theCasma, century, Moche, and Moquegua Valleys (Magilligan and

opened, could itnot be stuffedback in again? A number of possibilities may be raised, though they
are somewhat

why,

once

the Pandora's

box

of warfare

was

First of all, local sinchismay have profitedfrom derived their war leadership, preferred status from and must have found an extended climate of hos more conducive to theirends thanpeace. The tility chronicles hint at thisby attributing wars and war mongering in theTiticaca Basin toparticular lead
encouraging war, as in the Wanka area. Sinchis

conjectural

at this point.

Miraflores flooding may also indicatea drought cri

sis at some time soon after A.D. 1300.

Goldstein 2001; Reycraft 2000; Satterlee et al. 2000;Wells 1990). Because ENSO events are asso ciated with severe drought in the altiplano, the

While theclimate sequence in the south-central Andes is still rather poorly understood, multiple lines of evidence suggest dry periods in the later
thirteenth and fourteenth centuries, a situation that

ers, and by suggesting that the rewards of victory went firstand foremost to sinchis, not to fighting
forces or whole communities. For instance, Cieza

would have resulted in frequent yet unpredictable crop failure.Climate change thereforeappears to be a persuasive explanation for the escalation of

Here, warfare appears to solidify the hierarchical relationshipbetween a Colla sinchi and his people warfare in the fourteenth is the more century. This through the redistributionof spoils. because of large-scale in correlations Second, thepotential of fortificationsto them convincing the timingof fortification and defensive settlement selves create durable landscapes ofwar should not
across the southern and central Andes. Neverthe pukaras con less, the construction and and use of Colla

states the Colla paramount lord Zapana and his heirs "won many spoils inbattles,which theygave to the people of the district" (1984:279 [1553]).

tinued into the 1400s, when

accumulation, caya all suggest must oxygen

lake cores,
from Quelc more

ice

be underestimated. Fortifications in balkanized regions such as thisone, because theygive strength

to the defenders, make conquest, regional consol

tiful. After
fort-building

its initial impetus, warfare

have continued for other

isotopes that precipitation was

plen

and

reasons.

Does this sort of environmental explanation verge on environmental determinism, as Erickson (1999,2000) cautions? Erickson's emphasis on the resilience and ingenuityofAndeans inadapting to
their dynamic, sailable. and But unpredictable violent or actions, attacking environment too?raiding other is unas for stores to

idation, and the achievement of lasting peace dif ficult.Colla pukaras must have been particularly effectivebecause two of the most useful offensive tacticsagainst fortifications?prolonged sieges and
surprise attacks?were

livestock,

communities

pukaras suggest prolonged sieges were beyond the logistical capabilities of Colla societies, and the commanding views of pukaras in the treeless ter rainof thealtiplanowould have prevented most sur
prise attacks by large war and may parties. have Hence, encouraged pukaras local thwarted conquest

sible.The lackof year-round water sources onmost

apparently

rare or

impos

drive themfromcoveted lands?had

a place in the

364

LATIN AMERICAN

ANTIQUITY

[Vol. 19, No. 4, 2008

Colla pukaras: radiocarbon dates

Drought periods, from Quelccaya

6180

AZ4 (AA54251383?59BP) house I

AZ4 house II (AA54233 955?35BP)_.JIJ^A-JLL J*_ AZ4wall(M54220813?81BP)_-1 LJJmMJ^

IL_

JM|m^^^_

_. Jk L2(AA54245671?34BP) J^_

_ I L2(AA54246766?46BP)_"tH^

L2(AA12874580?60BP)_. JBA>__

L6(combineddates) !|f
2 S4unit(combined _ _ dates)
1 S4 unit (combined dates) _ _

j
1^_

|||j

!Sjj^^HBlEn

^|

CA2 (combined dates)

AS5(combineddates) _

: jg IffeS k"
_ ^MJBk^.. _ .

_-Am^__

I tsM j _JpjB^:-: H 1j jB?;;

AS1 (AA54241631?34BP) house ^;

PKP7 wall(combined dates) _ . _jpl|A..

H# | |??J AS1wall(AA54217427?70BP)
_ ...

!p | 2?| I _jL^^

il 11<W'XM

I I 4P^ f'^1^^?
H^*j II \\ ^^::m^m j \'-JM. -!-S

AZ1 (combineddates)
_. J10wall(AA54221516?41BP) *.

|
^M ._

-llll

| ^

J11 (combined wall dates)

AS3wall(AA54219502?35BP) _?

(to_jNL_

| :;;;1|VUM ?? HMpj

N2wall(AA54226709?34BP) j|? 1 |1| j || | fef|!:; j I :$;. SjjS;

N2wall(AA54225486?38BP)|j||; :|||||||

N2 1(AA64235510?34BP) jJf ||i | j| unit lil! |:||; | J|p | l|||i

[W[mMM N2unit2(combineddates) |:||M fill| ||||j
PKP2wall(AA54227455?34BP) _Mn_

jj| J j||| j jj |

^ IjM^ ;;; ^

^j

11 ;^-|fe;^

j ?|? |j

11 W&^?:u:^i !!jl|lfcj:2'

900

1000 1100 1200 1300 1400 1500 1600 Date AD

between pukara dates (black shapes) and peak droughts (vertical gray bars), as indicated by Figure 11. Comparison 8180, core 1 (Thompson 1992). Droughts are defined as years inwhich the 10 yr running mean > avg + la for Quelccaya whole core sequence.

Arkush]

WAR, CHRONOLOGY,

AND CAUSALITY

INTHE TITICACA

BASIN

365

groups to remain semi-independent and potentially aggressive. Perhaps the best indication of this dimension of pukaras is their reuse in theColla
uprising

tencemay have encouraged thedecision to rebel. Justas Erickson (2000) considers theaccumulated patrimony of raised fields, hillside terraces, improved qochas, and managed bofedales to be

against

the Incas.

Indeed,

their simple

exis

pie negotiated local intergrouprelationships with only limited conflict for at least two centuries in wake of a disintegration of state authority that the must have significantly reordered these relation It is highly probable thatdrought and attendant resource stressplayed a significant part in theesca lation of war in the lateLate IntermediatePeriod. Resource stress could have fostered violent com petition for limited arable land, and encouraged raids on stored crops; it may have indirectlyled to livestock rustlingas people became more relianton camelids, or perhaps itsimply caused greater social
ships.

part of Titicaca Basin Andeans' "landscape capi tal," so too pukaras, once built, became a kind of landscape capital, already inplace for thedefense of local communities against whatever enemies they might have ormake.

Third, the logic of segmentary social organiza tion, inwhich thekillings of individuals are seen as group offences demanding group vengeance (Kelly 2000), may have encouraged vendettas and facilitated the conscription of related groups into
larger-scale conflicts. Persistent histories of antag

friction between neighboring communities inhard times. Nevertheless, in the Titicaca Basin, fort building continued after the droughts eased, sug gesting thatwarfare outlived this impetus and generated itsown momentum. Good timesdid not
widespread war. The causes of continued war are

histories of conflict between Collas, Lupacas, and

Canas, reused

onism were materialized on the landscape in clus tersof defensive settlementand buffer zones, and the largest-scale enmitieswere remembered inoral
eventually recorded by the chroniclers. the memories Just

result in peace anymore than collapse resulted in

difficult to pinpoint securely, but power-hungry and the cycles of warlords, durable fortifications,

as Late Intermediate Period fortificationswere

by

expediently: shortly after the Spanish conquest, Lupacas took theopportunityofweakened central control to attack Colla settlements (Hemming 1970:242). The escalated violence of thefourteenth
century left a wake of bloodshed far beyond its ini tial impetus.

hostilities remained long after theLate Intermedi ate Period had ended, and could be stirred again up

later generations,

revenge encouraged by segmentary organization Collas inperiodic wars for may have embroiled the therestof the Late Intermediate Period and beyond.
I am grateful to the anonymous review Acknowledgments. ers, whose comments were very helpful in revising this arti cle, as well as Charles Stanish, Mark Allen, and Aimee Plourde, who gave insightful comments on an earlier draft. Funds for the radiocarbon dating of organic materials were Dissertation by a NSF provided Improvement Grant.

of past

was supported by an International Studies and Programs fellowship from UCLA, and by research grants from theDepartment of Anthropology, the Friends of and the Latin American Center at UCLA. I Archaeology, would like to thank the Instituto Nacional de Cultura of Peru Fieldwork Overseas for permission to work in theColla area, as well as the peo ple of the local INC office in Puno for their gracious aid. I thank Chip Stanish, Edmundo de la Vega, Cecilia Chavez, members of Programa Collasuyo, and numerous crew mem bers and colleagues for their support.

Conclusions
Colla warfare can be best understood from a mul

ticausal framework of constraints and incentives operating at different scales. The chronology of fortificationin the northernTiticaca Basin corre lateswith a number of other regions in the central

References Cited
Abbott, Mark B., Michael W. Binford, Mark Brenner, and Kerry R. Kelts 1997 A 3500 l4C yr High-Resolution Record ofWater Level Changes in Lake Titicaca, Bolivia-Peru. Quater nary Research (47): 169-180. Albarracin-Jordan, Juan 1996 Tiwanaku. Arqueologia Regional y Dinamica Seg mentaria. Editores Plural, La Paz. Albo, Xavier 1987 Comments on "Lenguas y pueblos altiplanicos en

Middle Horizon states set the stage for limitedfac tionalconflict in theearlyLate IntermediatePeriod, other factors contributed to a horizon of violence
centuries later. This sequence suggests that warfare

and southernAndean highlands. Together, these cases demonstrate thatwhile the collapse of the

is a possible, but not an inevitable, result of state

collapse. In the central and southern Andes, peo

366LATIN AMERICAN torno al Siglo XVI" 5(2):375-377. Allen, Mark W. 2006 by Alfredo Torero. Revista Andina

ANTIQUITY

[Vol.

19, No. 4, 2008

Arkush, Elizabeth 2005 Colla fortified sites:Warfare and regional power in the Late Prehispanic Titicaca Basin, Peru. Unpublished Ph.D. dissertation, Department ofAnthropology, Univer sity of California, Los Angeles. 2009 Warfare, Space, and Identity in the South-central Constraints and Choices. In Warfare in Cultural Context: Practice, Agency, and theArchaeology of Con flict, edited by Axel Nielsen andWilliam Walker. Amerind Foundation Advanced Seminar Series, University of Ari zona Press, Tucson, in press. Andes: del Sur, Tacna,

Allen, Mark W., and Elizabeth Arkush 2006 Introduction: Archaeology and the Study ofWar. In The Archaeology ofWarfare: Prehistories of Raiding and Conquest, edited by Elizabeth Arkush andMark Allen, pp. 1-22. University of Florida Press, Gainesville.

Maori Warfare: Toa, Pa, and Pu. Transformations in In The Archaeology ofWarfare: Prehistories of Raiding and Conquest, edited by Elizabeth Arkush andMark Allen, pp. 184-213. University of Florida Press, Gainesville.

1997 Climate Variation and theRise and Fall of anAndean Civilization. Quaternary Research 47:235-248. Bonnier, Elisabeth 1978 Habitat en village, a l'epoque prehispanique dans le bassin Shaka-Palcamayo. Bulletin, Institutfrangais d'e tudes andines 7( l-2):49-71. Bonta, Bruce D. 1997 Cooperation and Conflict inPeaceful Societies. Psy chological Bulletin 121 (2): 299-320.

Therese Bouysse-Cassagne, 1987 La identidad Aymara. Hisbol, La Paz. Browman, David 1994 Titicaca Basin Archaeolonguistics: Uru, Pukina and A.D. 750-1450. World Archaeology Aymara, 26(2):235-251. Brubaker, Rogers, and David Laitin 1998 Ethnic and Nationalist Violence.

Sociology 24:423-452. Cabello Balboa, Miguel 1951 [1586] Misceldnea Antdrtica: Una Historia del Peru Antiguo. Universidad Nacional Mayor de San Marcos, Lima. Cal away, Michael J. 2005 Ice-cores, Sediments

Annual Review of

Ayca Gallegos, Oscar 1995 Sillustani. Instituto de Arqueologia Peru.

Baker, Paul A., Geoffrey O. Seltzer, Sherilyn C. Fritz, Robert B. Dunbar, Matthew J.Grove, Pedro M. Tapia, Scott L. Cross, Harold D. Rowe, and James P. Broda 2001 The History of South American Tropical Precipitation for the Past 25,000 Years. Science 291:640-643. Bamforth, Douglas B. 2006 Climate, Chronology, and theCourse ofWar in the Middle Region of theNorth American Great In The Archaeology of Warfare: Prehistories of Raiding and Conquest, edited by Elizabeth Arkush and Mark Allen, pp. 66-100. University of Florida Press, Missouri Plains.

and Civilisation Collapse: A Tale from Lake Titicaca. Antiquity Cautionary 79(306):778-791. Cerron-Palomino, Rodolfo In Wari y del Aimara. El origen centroandino 2000 Tiwanaku: Modelos y Evidencias, Primera Parte. Boletin de Arqueologia PUCP 4, edited by Peter Kaulicke and William H. Isbell, pp. 131-142. Pontificia Universidad Catolica del Peru, Lima. Chagnon, Napoleon A. 1968 Yanomamo Social Organization andWarfare. InWar: the Anthropology ofArmed Conflict and Aggression, edited byMorton Fried, Marvin Harris and Robert Murphy, pp. 110-159. Natural History Press, Garden City, NY. 1988 Life Histories, Blood Revenge andWarfare ina Tribal

Gainesville. Barreda Murillo, Luis en Nunoa Kollao 1958 Sitios Archeologicos (Melgar Bachelor's thesis, Universidad Puno). Unpublished Nacional de Cuzco. Bauer, Brian 1992 The Development of theInca State. University ofTexas Press, Austin. 2004 Ancient Cuzco. University of Texas Press, Austin. Bauer, Brian, and R. Alan Covey 2002 Processes of State Formation in the Inca Heartland (Cuzco, Peru). American Anthropologist 104(3): 846-865. Bennett, Wendell C. 1933 Archaeological 33(2): 163-174. Hikes

in theAndes. Natural History in the Titicaca Basin.

Population. Science 239:985-992. Cieza de Leon, Pedro 1984 [1553] Cronica del Peru: Primera Parte. Pontifica Universidad Catolica del Peru, Lima. 1985 [1553] Cronica del Peru: Segunda Parte. Pontifica Universidad Catolica del Peru, Lima. Cobo, Bernabe 1979 [ 1653] History of the Inca Empire. Translated by R. Hamilton. University of Texas Press, Austin. Covey, R. Alan 2008 Multiregional Perspectives on the Archaeology of the Andes during the Late Intermediate Period (c. A.D. Research Journal of Archaeological 1000-1400). 16:287-338. D'Altroy, Terence N., and Christine Hastorf (editors). 2001 Empire and Domestic Economy. Plenum, New York. Daly, Martin, and Margo Wilson 1988 Homicide. Aldine de Gruyter, New York, de laVega M., Edmundo 1990 Estudio arqueologico de Pucaras o poblados amural ladas de cumbre en territorioLupaqa: El caso de Pucara Juli.Unpublished Bachelor's thesis,Universidad Catolica

Bertonio, Ludovico 1986 [1612] Vocabulario de la lengua Aymara. Facsimile ed. CERES, IFEA, and MUSEF, La Paz. Betanzos, Juan de 1996 [1551-7] Narrative of the Incas. Translated by D. Buchanan R. Hamilton. University of Texas Press, Austin. Billman, Brian R., Patricia M. Lambert, and Banks L. Leonard Mesa Verde 2000 Cannibalism, Warfare, and Drought in the

1950 Cultural Unity and Disunity American Antiquity 16:89-98.

Region during the 12thCentury A.D. American Antiquity 65:145-178. W. Binford, Michael W., Alan L. Kolata, Mark Brenner, John Janusek,Matthew T. Seddon, Mark Abbott, and Jason H. Curtis

Santa Maria, Arequipa. de laVega M., Edmundo, Kirk L. Frye, and Tiffiny Tung In The Cave Burial from Molino-Chilacachi. 2005 Advances inTiticaca Basin Archaeology, edited by Charles Stanish, Amanda B. Cohen and Mark S. Aldenderfer, pp. 173-184. Cotsen Institute of Archaeology, UCLA, Los Angeles.

Arkush]

WAR, CHRONOLOGY,

AND CAUSALITY

INTHE TITICACA BASIN

367

Desjarlais, Robert, and Arthur Kleinman in theNew World Dis 1994 Violence and Demoralization order. Anthropology Today 10(5):9?12. Diez de San Miguel, Garci 1964 [ 1567] Visita hecha a la Provincia de Chucuito. Casa de la Cultura del Peru, Lima. Earle, Timothy, Terence D'Altroy, and Catherine LeBlanc 1980 Arqueologia Regional de los Periodos Prehispanicos Tardios en elMantaro. In El Hombre y la Cultura Andina

DeMarrais, Elizabeth 1997 Materialization, Ideology, ment of Centralized Authority Polities of theValle Calchaqui, Ph.D. dissertation, Department sity of California, Los Angeles.

and Power: The develop among the pre-Hispanic Argentina. Unpublished ofAnthropology, Univer

fer,pp. 173-184. Cotsen Institute ofArchaeology, UCLA, Los Angeles. Fuentes G., Cleta J. 1991 Arqueologia del Distrito de Asillo (Prospeccion). thesis,Universidad Nacional de Unpublished Bachelor's San Antonio Abad del Cusco, Cusco, Peru. Garreaud, Rene D., and Patricio Aceituno 2001 Interannual Rainfall Variability over the South Amer icanAltiplano. Journal of Climate 14:2779-2789. Gat, Azar The Causes and Origins of "Primitive Warfare": to Ferguson. Anthropological Reply Quarterly 73(3): 165-168. 2000b The Human Motivational Complex: Evolutionary of Hunter-Gatherer Fighting. Theory and the Causes 73(2):74-88. Anthropological Quarterly 73(l):20-34, Gleditsch, Nils Petter (editor) 1997 Conflict and the Environment. Kluwer, Dordrecht, Netherlands. 2000a Haas, Jonathan 1999 The Origins ofWar and Ethnic Violence. In Ancient Warfare: Archaeological Perspectives, edited by JohnCar man and Anthony Harding, pp. 11-24. Sutton, Stroud, Great Britain. Haas, Jonathan, andWinifred Creamer 1993 Stress and Warfare among the Kayenta Anasazi of the 13th Century AD. Fieldiana, Anthropology n.s. No. 21. Field Museum of Natural History, Chicago. Harris, Marvin 1974 Cows, Pigs, Wars, andWitches: The Riddles of Cul ture.Vintage, New York. Hastenrath, Stefan, Dierk Polzin, and Bernard Francou 2004 Circulation Variability Reflected in Ice Core and Lake

(III Congreso Peruano), Vol. II, pp. 641-670. Editorial Ramiro Matos, Lima. Earle, Timothy, Terence D'Altroy, Christine Hastorf, Catherine Scott, Cathy Costin, Glenn Russell, and Elsie Sandefur 1987 Archaeological Field Research in the Upper Mantaro, Peru,

1982-1983: Investigations of Inka Expansion and Exchange. Insititute of Archaeology, UCLA, Los Ange les.

Ember, Carol R., and Melvin Ember 1992 Resource Unpredictability, Mistrust, and War: A Cross-Cultural Study. Journal of Conflict Resolution 36(2):242-262. Erickson, Clark 1999 Neo-environmental Determinism and Agrarian 'Col lapse' inAndean Prehistory. Antiquity 73:634-642. 2000 The Lake Titicaca Basin: A Precolumbian Built Land scape. In Imperfect Balance: Landscape Transformations in the Precolumbian Americas,

1997 Political Centralization in theAltiplano Period in the Soutwestern Titicaca Basin (Appendix 2). InArchaeolog ical Survey in the Juli-Desaguadero Region of Lake Titi caca Basin, Southern Peru, edited by Charles Stanish. Fieldiana, Anthropology n.s. No. 29. Field Museum of Natural History, Chicago. Frye, Kirk L., and Edmundo de laVega M. 2005 The Altiplano Period in the Titicaca Basin. In Advances in theArchaeology of theTiticaca Basin, edited by Charles Stanish, Amanda Cohen and Mark Aldender

sity Press, Cambridge. 2001 Materialist, Cultural, and Biological Theories onWhy Yanomami Make War. Anthropological Theory 1(1): 99-116. Frye, Kirk L.

de la historia militar inca. Allpanchis phutur inqa 16:171-186. 1987 Los Incas. Amaru Editores, Lima. Evans-Pritchard, Edward E. 1940 The Nuer of the southern Sudan. InAfrican Political Systems, edited by Meyer Fortes and Edward E. Evans Pritchard. Oxford University Press, London. Farfan L., Carlos 1995 Asentamientos prehispanicos de la cuenca alta del Chillon. Gaceta Arqueological Andina (24):31 -61. Ferguson, R. Brian 1984 Introduction: StudyingWar. InWarfare, Culture, and Environment, edited by R. Brian Ferguson, pp. 1-81. Aca demic Press, Orlando, Florida. 1990 Explaining War. In The Anthropology of Warfare, edited by Jonathan Haas, pp. 26-55. Cambridge Univer

edited by D. Lentz, pp. 311-356. Columbia University Press, New York. Espinoza S.,Waldemar 1980 Acerca

Records of the Southern Tropical Andes. Climatic Change 64(3):361-375. Hastorf, Christine A, Elsie Sandefur, Timothy K Earle, H. E. Wright. Lisa LeCount, and Glenn Russell 1989 Settlement Archaeology in the Jauja Region of Peru:

Evidence from theEarly Intermediate Period through the Late Intermediate Period: A Report on the 1986 Field Sea son. Andean Past 2:81-129.

Hefferman, Ken 1996 Limatambo: Archaeology, History and theRegional Societies of Inca Cusco. BAR International Series No. 644. British Archaeological Reports, Oxford. Hoffmann, Georg, Edson Ramirez, Jean-Denis Taupin, Bernard

30(4): 1179-1182. Homer-Dixon, Thomas F. 1999 Environment, Scarcity, and Violence. Princeton Uni versity Press, Princeton, New Jersey. Hyslop, John 1976 An Archaeological Investigation of theLupaqa King dom and its Origins. Unpublished Ph.D. dissertation, Department of Anthropology, Columbia University. Isko, Xavier 1992 La doble frontera: Ecologia, politica y ritual en el Altiplano Central. Hisbol/Ceres, La Paz. Jackson, Robert H. 1990 Quasi-States: Sovereignty, International Relations, and theThird World. Cambridge University Press, Cam bridge.

Francou, Pierre Ribstein, Robert Delmas, Hans Diirr, Robert Gallaire, Jefferson Simoes, Ulrich Schotterer, Michel Stievenard, and Martin Werner 2003 Coherent Isotope History of Andean Ice Cores over the Last Century. Geophysical Research Letters

368LATIN AMERICAN Jakobsen, Jan, J. Balslev Jorgensen, L. Kempfner Jorgensen, and Inge Schjellerup 1986-1987 "Cazadores de cabezas" en sitios pre-inca de Chachapoyas, Amazonas. Revista del Museo Nacional 48:139-186. W. Janusek, John 2004 Identity and Power in theAncient Andes: Tiwanaku Cities throughTime. Routledge, New York. Johnson, liana 2003 The Changing Nature of Political Power in theTiti caca Basin: The View from Paucarcolla. Unpublished Mas ter's thesis, Department of Anthropology, University of California, Los Angeles. Jones, Terry L., Gary M. Brown, L. Mark Raab, Janet L. J. Kennett, McVickar, W. Geoffrey Spaulding, Douglas 1999

ANTIQUITY

[Vol.

19, No. 4, 2008

Kusimba, Chapurukha 2006 Slavery and Warfare inAfrican Chiefdoms. In The Archaeology of Warfare: Prehistories ofRaiding and Con quest, edited by Elizabeth Arkush and Mark Allen, pp. 214-252. University Press of Florida, Gainesville. Lambert, Patricia M. in Prehistoric Hunter-gatherer 1997 Patterns of Violence Societies of Coastal Southern California. In Troubled Times: Violence and Warfare in thePast, edited by Debra L. Martin and David W. Frayer, pp. 77-110. Overseas Publ. Assoc, Amsterdam. 2002 The Archaeology ofWar: A North American Per of Archaeological Research

Journal spective. 10(3):207-241. Larson, Lewis H.

Climatic Anomaly. Current Anthropology 40(2): 137-170. Julien, Catherine A View of Inca Rule from theLake Titi 1983 Hatunqolla: caca Region. Series Publications in Anthropology No. 15. University of California, Berkeley. Julien,Daniel 1988 Ancient Cuismancu: Settlement and Cultural Dynam ics in the Cajamarca Region of theNorth Highlands of Medieval 1532. Unpublished Ph.D. disserta Peru, 200 B.C.-A.D. tion, Department of Anthropology, University of Texas, Austin. 1993 Late Pre-Inkaic Ethnic Groups inHighland Peru: An Model of the Political Region. Latin American

Andrew York, and Phillip L. Walker Demo Environmental Imperatives Reconsidered: graphic Crises inWestern North America during the

1972 Functional Considerations ofWarfare in the South east During theMississippi Period. American Antiquity 37:383-392. Lavallee, Daniele, and Michele Julien 1973 Les etablissements Asto a Vepoque prehispanique. Travaux de lTnstitut Francais d'Etudes Andines No. 15, Lima. LeBlanc, Catherine J. 1981 Late Prehispanic Huanca Settlement Patterns in the Yanamarca Valley, Peru. Unpublished Ph.D. dissertation, University of California, Los Angeles. LeBlanc, Steven A. American Southwest. Uni 1999 Prehistoric Warfare in the versity of Utah Press, Salt Lake City. 2006 Warfare and theDevelopment of Social Complexity: and Environmental Factors. In The Some Demographic

Archaeological-ethnohistorical Geography of the Cajamarca Antiquity 4:246-273. Keeley, Lawrence H. 1996 War Before Civilization. New York.

Archaeology of Warfare: Prehistories ofRaiding and Con quest, edited by Elizabeth Arkush and Mark Allen, pp. 437-468. University Press of Florida, Gainesville. Steven A., with Katherine Register Constant Battles: The Myth of the Peaceful, Noble Savage. St. Martin's Press, New York.

Oxford University Press,

LeBlanc, 2003

Kelly, Raymond 2000 Warless Societies and theOrigin ofWar. University ofMichigan Press, Ann Arbor. Kendall, Ann 1996 An Archaeological Perspective forLate Intermediate in theCuzco Region. Journal of the Period Development Steward Anthropological Society 24( 1-2): 121-156. Kidder II, Alfred Northern Lake Titicaca Basin. 1943 Some Early Sites in the Papers of thePeabody Museum ofAmerican Archaeology and Ethnology No. 27, Harvard University, Cambridge, Massachusetts. Portrait of an Andean Civilization.

Lecoq, Patrice 1997 Patron de asentamiento, estilos ceramicos y grupos etnicos: El ejemplo de la region Intersalar en Bolivia. In Saberes yMemorias en los Andes: InMemoriam Thierry pp. 59-89. Saignes, edited by Therese Bouysse-Cassagne, Paris. IHEAL-IFEA, Lecoq, Patrice, and Ricardo Cespedes 1997 Panorama archeologique des zones meridionales de Bolivie (sud-est de Potosf). Bulletin de lTnstitutFrancais d'Etudes Andines 26(1):21-61.

Kolata, Alan 1993 The Tiwanaku:

Blackwell, Cambridge, Massachusetts. Kolb, Michael J.

Krzanowski, Andresz Alto Chi 1977 Yuraccama, the Settlement Complex in the cama Region (Northern Peru). In Polish Contributions in New World Archaeology, edited by Janusz Kozlowski, pp. 29-58. Zakland Narodwy imOsolinskich, Krakow. 1983 The Cultural Chronology of theNorthern Andes of Peru (the Huamachuco-Quiruvilca-Otuzco Region). Sociedad Polaca de Estudios Latinoamericanos, Comision Warsaw. Andina, Documentos de Trabajo No. 1,

1991 Social Power, Chiefly Authority, and Ceremonial Architecture in an Island Polity, Maui, Hawaii. Unpub lished Ph.D. dissertation, Department of Anthropology, University of California, Los Angeles.

Lekson, Stephen World. American 2002 War in the Southwest, War in the Antiquity 67:607-624. Llagostera M., Agustfn, andMaria Costa J. 1999 Patrones de asentamiento en la epoca agroalfarera de San Pedro de Atacama (norte de Chile). Estudios Ata camenos 11:30-34. Magilligan, Francis J., and Paul Goldstein 2001 El Nino Floods and Culture Change: A Late Holocene

Flood History for theRio Moquegua, Southern Peru. Geol ogy 29(5):431-434. Mantha, Alexis 2006 Late Prehispanic Social Complexity in theRapayan Valley, Upper Maranon Drainage, Central Andes of Peru. In La Complejidad Social en la Sierra de Ancash, edited by Alexander Herrera, Carolina Orsini, and Kevin Lane, pp. 35-62. PUNKU Centro de Investigation Andina, Lima. Matos M., Ramiro 1999 The "Senorios" in the Sierra and theCentral Coast

Arkush]

WAR, CHRONOLOGY,

AND CAUSALITY 2007

INTHE TITICACA

BASIN

369

Meddens, Frank 1999 La secuencia cultural de laCuenca de Chicha Soras. InXII Congreso Peruano del Hombre y la Cultura And ina,Vol. 2, edited by Ismael Perez, Walter Aguilar S., and Medardo Purizaga V, pp. 200-210. Universidad Nacional San Cristobal de Huamanga, Ayacucho, Peru. Meggitt, Mervyn 1977 Blood is theirArgument. Mayfield, Palo Alto, Cali fornia. Melice, Jean-Luc, and Pascal Roucou 1998 Decadal Time Scale Variability Recorded in theQuel ccaya Summit Ice Core _,80 Isotopic Ratio Series and its Relation with the Sea Surface Temperature. Climate

In The Inca World: The Development of Pre-Columbian Peru, AD 1000-1534, edited by Laura Laurencich Minelli. University of Oklahoma Press, Norman.

Late Chachapoya Population Structure prior to Inka Conquest. American Journal of Physical Anthropology 131(3):334-342. W. Verano Nystrom, Kenneth C, and John 2003 Warriors of the clouds? Inference and interpretation of trauma fromChachapoya, Peru. Paper presented at the 73rdAnnual Meeting of the American Association of Phys ical Anthropology, Tempe, Arizona. 2006 Orlove, Benjamin S., JohnC. H. Chiang, and Mark A. Cane 2000 Forecasting Andean Rainfall and Crop Yield from the Influence of El Nino on Pleiades Visibility. Nature 403:68-71.

Times of Plenty, Times of Less: Last-Millennium Societal Disruption in the Pacific Basin. Human Ecology 35(4):385-401. Nystrom, Kenneth C.

Dynamics 14(2): 117-132. Milner, George E. 1999 Warfare in Prehistoric and Early Historic Eastern North America. Journal of Archaeological Research 7(2):105-151. 2007 Warfare, Population, and Food Production inPrehis toric Eastern North America. InNorth American Indige nous Warfare and Ritual Violence, edited by Richard

Ortloff, Charles, and Alan Kolata 1993 Climate and Collapse: Agroecological Perspectives on the Decline of the Tiwanaku State. Journal of Archaeo logical Science 20:195-221. Otterbein, Keith 1970 The Evolution of War. HRAF, New Haven, Con necticut. 2004 How War Began. Texas A&M University Anthropol ogy Series, College Station. Owen, Bruce 1995 Warfare and Engineering, Ostentation and Status in the Late Intermediate Period Osmore Drainage. Paper pre sented at the 60th Annual Meeting of the Society for Amer ican Archaeology, Minneapolis. 2005 Distant Colonies and Explosive Collapse: The Two Stages of theTiwanaku Diaspora in theOsmore Drainage. Latin American Antiquity 16:45-80.

Chacon and Ruben Mendoza, Arizona Press, Tucson. Moseley, Michael E.

pp. 182-201. University of

1989 Fortificaciones prehispanicas y evolucion de tacticas militares en el valle de Moquegua. In Trabajos Arque Peru, Vol. 1, edited by Luis K. ologicos en Moquegua, Watanabe, Michael E. Moseley, and Fernando Cabieses, pp. 237-252. Programa Contisuyo del Museo Peruano de Ciencias de la Salud y Southern Peru Copper Corporation, Lima.

2002 Modeling Protracted Drought, Collateral Natural Dis Andes. InCatastrophe aster, and Human Responses in the & Culture: The Anthropology of Disaster, edited by Susanna M. Hoffman and Anthony Oliver-Smith, pp. 187-212. School of American Research Press, Santa Fe, New Mexico. Murua, Martin de 1986 [1613] Historia General del Peru. Historia 16, Madrid. Narvaez V, Luis A. 1987 Kuelap: Una ciudad fortificada en los Andes nor orientales de Amazonas, Peru. In Arquitectura y Arque ologia: Pasado y Futuro de la Construccion en el Peru, edited by Victor Rangel F, pp. 115-142. Universidad de

Parsons, Jeffrey R., and Charles M. Hastings 1988 The Late Intermediate Period. In Peruvian Prehis tory,edited by Richard W. Keatinge, pp. 190-229. Cam

Chiclayo, Chiclayo, Peru. Neira A., Maximo 1967 Informe preliminar de las investigaciones arqueolog icas en el Departamento de Puno. Anales del Insituto de Estudios Socio Economicos 1(1): 107-164, Puno, Peru. Nielsen, Axel E. 2001 Evolucion social en Quebrada de Humahuaca (AD InHistoria Argentina Prehispdnica I, edited 700-1536). by Eduardo Berberian and Axel Nielsen, pp. 171-264. Editorial Brujas, Cordoba, Argentina. 2002 Asentamientos, conflicto y cambio social en el alti plano de Lipez (Potosi). Revista Espahola de Antropologia Americana 32:179-205.

bridge University Press, Cambridge. R., Charles M. Hastings, and Ramiro Matos Parsons, Jeffrey M. 1997 Rebuilding the State in Highland Peru: Herder cultivator Interaction during theLate Intermediate Period in the Tarama-Chinchaycocha Region. Latin American Antiquity 8:317-341. 2000 Prehispanic Settlement Patterns in the Upper Man taro and Tarma Drainages, Junin, Peru. Memoirs of the of Anthropology No. 34. 2 vols. Museum Museum of Anthropology, University ofMichigan, Ann Arbor. Parssinen, Martti H. 2005 Caquiaviri y la Provincia de Pacasa (Bolivia), desde el Alto Formativo hasta la Conquista Espanola (1-1533). CIMA, La Paz. Parssinen, Martti H., and Ari Siiriainen 1997 Inka-style Ceramics and their Chronological Rela tionship to the Inka Expansion in the Southern Lake Titi caca Area (Bolivia). Latin American Antiquity 8:255-271. Petersen, Kenneth L. 1988 Climate and the Dolores River Anasazi. University of Utah Press, Salt Lake City. Piatt, Tristan 1986 Mirrors andMaize: The Concept ofYanantin among the Macha of Bolivia. In Anthropological History of Andean Polities, edited by John Murra, Nathan Wachtel, and Jacques Revel, pp. 228-259. Cambridge University Press. Cambridge. 1987 Entre ch'axwa y muxsa. Para una historia del pen samiento politico Aymara. In Tres reflexiones sobre elpen

Nunez, Lautaro, and Tom Dillehay 1978 Movilidad giratoria, armonia social y desarollo en los Andes meridionales: Patrones de trafico e interaction economica. Universidad del Norte, Antofagasta, Chile. Nunn, Patrick, Rosalind Hunter-Anderson, Mike Carson, Frank Thomas, Sean Ulm, and Michael Rowland

370LATIN AMERICAN samiento andino, edited by Therese Bouysse-Cassagne, Olivia Harris, Tristan Piatt, and Veronica Cereceda, pp. 61-132. Hisbol, La Paz. Plourde, Aimee, and Edmundo de laVega M. 2001 Reconocimiento Sistematico e Intensivo del Valle del Rio Huancane-Putina. Report National de Cultura, Lima. Polo de Ondegardo, Juan submitted to the Instituto

ANTIQUITY

[Vol.

19, No. 4, 2008

tion,Universidad Nacional Mayor de San Marcos, Lima. Sahlins, Marshall 1961 The Segmentary Lineage: An Organization of Preda toryExpansion. American Anthropologist 63:322-345. 1968 Tribesmen. Prentice Hall, Eaglewood Cliffs, New Jer sey. Santa Cruz Pachacuti Yamqui Salcamayhua, Juan de 1993 [1613] Relacion de antiguedades deste reynodel Piru. Travaux de lTnstitut Francais d'Etudes Andines No. 74. InstitutFrancais d'Etudes Andines, Lima. Santoro, Calogero M., Alvaro Romero G., Vivien G. Standen, and Amador Torres 2004 Continuidad y cambio en las comunidades locales, Perfodos Intermedio Tardio y Tardio, valles occidentales del areaCentro Sur Andina. Chungard 36(supl.):235-247. Sarmiento de Gamboa, Pedro 1988 [1572] Historia de los Incas. Biblioteca de Viajeros /Polifemo, Madrid. Hispanicos 4. Miraguano Satterlee, Dennis R., Michael E. Moseley, and David K. Keefer 2000 The Miraflores El Nino Disaster: Convergent Cata in Southern strophes and Prehistoric Agrarian Change

1873 [1571] Narratives ofthe Rites and Laws of the Incas, edited and translated by Clements R. Markham. Hakluyt Society, London. Raab, L. Mark, and Daniel O. Larson 1997 Medieval Evolution Climatic Anomaly and Punctuated Cultural inCoastal Southern California. American Antiq

uity 62319-331. Rawls, Joseph M. 1979 An Analysis of Prehispanic Andean Warfare. Unpub lished Ph.D. dissertation, University of California, Los

No. 1385. British Archaeological Reports, Oxford. Reycraft, Richard M. to El Nino in South 2000 Long-term Human Response Coastal Peru. In Environmental Disaster and theArchae ology ofHuman Response, edited by Garth Bawden and Richard M. Reycraft, pp. 99-120. Maxwell Museum of Reyna, Stephen P. 1994 A Mode of Domination Approach toOrganized Vio lence. In Studying War: Anthropological Perspectives, edited by Stephen P. Reyna and R. E. Downs, pp. 29-65. Gordon and Breach, Langhorne, Pennsylvania. Rigsby, Catherine A., Paul A. Baker, and Mark S. Aldenderfer 2003 Fluvial History of theRio Have valley, Peru, and its to Climate and Human History. Palaeo Relationship geography, 194:165-185. Palaeoclimatology, Palaeoecology Anthropology, University of New Mexico, Albuquerque.

Angeles. Reindel, Markus 2005 Pinchango Alto: A Gold Miners' Settlement inPalpa, Southern Peru. In Wars and Conflicts in Prehispanic and the Mesoamerica Andes, edited by Peter Eeckhout and Genevieve Le Fort, pp. 90-98. BAR International Series

Peru. Andean Past 6:95-116. Schiappacasse, Virgilio, Victoria Castro, and Hans Niemeyer 1989 Los desarrollos regionales en el Norte Grande (1000 - 1400 d.C). In Culturas de Chile: Prehistoria, edited by and Hans Schiappacasse, Virgilio Jorge Hidalgo, Niemeyer, pp. 181-220. Editorial Andres Bello, Santiago. Schjellerup, Inge 1992 Patrones de asentamiento en las faldas orientales de In Estudions de losAndes de la region de Chachapoyas.

Peruana, edited by Duccio Bonavia, pp. Arqueologia 355-364. Fomciencias, Lima. of the in the Conquest Incas and Spaniards 1997 and Ethnohistorical Archaeological Chachapoyas: Research in the Northeastern Andes of Peru. Series B, Gothenburg Archaeological Thesis No. 7,Dept ofArchae ology, Gothenburg University. The National Museum of

Robarchek, Clayton, and Carole Robarchek 1998 Waorani: The Contexts of Violence and War. Har court Brace, Orlando, Florida. Roche, Michel A., Jacques Bourges, Jose Cortes, and Roger

Denmark, Copenhagen. Schultze, Carol into the Puno Bay Region of 2000 Tiwanaku Expansion Southern Peru. Paper presented at the 65th Annual Meet ings of the Society forAmerican Archaeology, Philadel phia. Seltzer, Geoffrey O., and Christine A. Hastorf 1990 Climatic Change and itsEffect on Agriculture in the Central Peruvian Andes. Journal of Field Archaeology 17(4):397-414. Service, Elman 1968 War and Our Contemporary Ancestors. In War: the Anthropology ofArmed Conflict and Aggression, edited by Morton Fried, Marvin Harris, and Robert Murphy, pp. 160-167. Natural History Press, Garden City, New York. Shaw, R. Paul, and Yuwa Wong 1989 Genetic Seeds ofWarfare: Evolution, Nationalism, and Patriotism. Unwin Hyman, Boston. Shimada, Izumi, Crystal B. Schaaf, Lonnie G. Thompson, and

emic, Dordrecht, Netherlands. Ross, Marc H. 1993 The Culture of Conflict. Yale University Press, New Haven. Rostworowski de Diez Canseco, Maria 1970 El repartamiento de dona Beatriz Coya, en el valle de Yucay. Historia y Cultura 4:153-268. Rowe, JohnH. 1942 Sitios historicos en la region de Pucara, Puno. Revista 10-11:66-75. del Instituto Arqueologico Andean Area. American 1945 Absolute Chronology in the Antiquity 3:265-284. 1985 Probanza de los Incas nietos de conquistadores. His torica 9(2): 193-245. Ruiz E., Arturo 1973 Las ruinas de Sillustani. Unpublished Ph.D. disserta

Mattos and Hydrology of the Lake Titicaca 1992 Climatology Basin. In Lake Titicaca: A Synthesis of Limnological Knowledge., edited by Claude Dejoux and Andre litis, pp. 63-88. Monographiae Biologicae, Vol. 68. Kluwer Acad

Ellen Mosley-Thompson 1991 Cultural Impacts of Severe Droughts in thePrehistoric Andes: Application of a 1,500-year Ice Core Precipitation Record. World Archaeology 22(2):247-270. Silva S., Jorge E. 1992 Patrones de asentamiento en el Valle de Chillon. In edited by Duccio Estudios de Arqueologia Peruana, Bonavia, pp. 393-403. Fomciencias, Lima. Simons, Anna 1995 Networks ofDissolution: Somalia Undone. Westview

Arkush]

WAR, CHRONOLOGY,

AND CAUSALITY

INTHE TITICACA

BASIN

371

Press, Boulder, Colorado. Snyder, Jack 2002 Anarchy and Culture: Insights from theAnthropol War. International Organization 56(l):7-45. ogy of Stanish, Charles Otora Val 1985 Post-Tiwanaku Regional Economies in the Ph.D. dissertation, ley, Southern Peru. Unpublished Department ofAnthropology, University of Chicago. 1992 Ancient Andean Political Economy. University of Texas Press, Austin. 2003 Ancient Titicaca: The Evolution of Complex Society in Southern Peru and Northern Bolivia. University of Cal

Bernard Francou 1998 A 25,000-year Tropical Climate History from Boli vian Ice Cores. Science 282:1858-1864. and Keith A. Thompson, Lonnie G., Ellen Mosley-Thompson,

Henderson in Tropical South 2000 Ice-Core Paleoclimate Records America since theLast Glacial Maximum. Journal ofQua ternary Science 15(4):377-394. Mary E. Thompson, Lonnie G., Ellen Mosley-Thompson, Davis, Ping-Nan Lin, Keith A. Henderson, and Tracy A. Mashiotta 2003 Tropical Glacier and Ice Core Evidence of Climate Change on Annual toMillenial Time Scales. Climatic Change 59(1-2): 137-155.

ifornia Press, Berkeley. Stanish, Charles, Kirk L. Frye, Edmundo de la Vega, and Matthew T. Seddon Western Titicaca Basin, 2005 Tiwanaku Expansion into the Peru. InAdvances inTiticaca Basin Archaeology I, edited by Charles Stanish, Amanda Cohen, andMark Aldender fer,pp. 103-114. Cotsen InstituteofArchaeology, UCLA, Los Angeles. Stanish, Charles, Cecilia Chavez J.,Edmundo De laVega M., and Adan Umire 2008 Archaeological survey in theArapa region of north ern Lake Titicaca. Working papers #4, Programa Colla suyu.Manuscript in possession of the author. Stanish, Charles, Edmundo de laVega, Lee Steadman, Cecilia Chavez, Kirk L. Frye, Luperio Onofre, Matthew T. Sed don, and Percy Calisaya 1997 Archaeological Survey in the Juli-Desaguadero Region of the Lake Titicaca Basin, Southern Peru. Fiel diana, Anthropology n.s. No. 29. Field Museum of Nat

Toledo, Francisco de 1975 [1570] Tasa de la Visita General de Francisco de Toledo. Universidad Nacional de San Marcos, Lima. Topic, JohnR., and Theresa L. Topic 1987 The Archaeological Investigation of Andean Mili In The Origins tarism: Some Cautionary Observations. and Development of the Andean State, edited by Jonathan Haas, Sheila Pozorski, and Thomas Pozorski, pp. 47-55. Cambridge University Press, Cambridge. Torero, Alfredo 1987 Lenguas y pueblos altiplanicos en torno al Siglo XVI. Revista Andina 5(2):329^105. 1992 Acerca de la familia linguistica Uruquilla. Revista Andina 19:171-91. Torres-Rouff, Christina, and Maria A. Costa in Prehistoric San Pedro de 2006 Interpersonal Violence Atacama, Chile: Behavioral Implications of Environmen tal Stress. American Journal of Physical Anthropology

ural History, Chicago. Tapia P., Felix 1978a Contribuciones al estudio de la cultura precolombina no. 16. Instituto en el Altiplano Peruano. Publication National de Arqueologia, La Paz. en Kacsili. Puma 1978b Investigaciones arqueologicas punku 12:7-37. en los 1985 Contribution a la Investigation Arqueologica en el Departamento Valles de Sandia y Carabaya, de Puno?Peru. Grupo de Arte Utaraya, Puno, Peru. Thayer, Bradley A. 2004 Darwin and International Relations: On the Evolu tionary Origins ofWar and Ethnic Conflict. University Press of Kentucky, Lexington. Thompson, Lonnie G. 1992 Quelccaya Ice Core Database. IGBP PAGES/World Data Center-A for Paleoclimatology Data Contribution Series # 92-008. NOAA/NGDC Pro Paleoclimatology gram, Boulder, Colorado. Data fromThompson et al. 1985. Electronic at available document, http://www.ncdc.noaa.gov/paleo/icecore/trop/quelc caya/quelccay a_data.html. John F. Thompson, Lonnie G., Ellen Mosley-Thompson, Bolzan, and Bruce R. Koci 1985 A 1500-year Record of Tropical Precipitation in Ice Cores 229:971-973. from the Quelccaya Ice Cap, Peru. Science

Gaceta Arqueologica Andina Qaracha (Ayacucho). 5(17): 17-26. van der Dennen, J. M. G. (editor) 1995 The Origin of War: The Evolution of a Male Coalitional Reproductive Strategy, Origin Press, Gronin gen, Netherlands. Vasquez, Emilio 1940 Itinerario arqueologico del Kollao. Revista delMuseo Nacional 9(1): 143-150. Vayda, Andrew P. 1976 War inEcological Perspective. Plenum, New York. W. Verano, John

DC). Estudios atacamehos (22):7?31. Valdez, Lidio M., and Cirilo Vivanco 1994 Arqueologia de la Cuenca del Qaracha, Ayacucho, Peru. Latin American Antiquity 5:144-157. Valdez, Lidio M., Cirilo Vivanco, and Casimiro Chavez 1990 Asentamientos Chanka en la cuenca del Pampas

130(l):60-70. Tschopik, Marion H. 1946 Some Notes on the Archaeology of the Department of Puno. Papers of the Peabody Museum of American Archaeology and Ethnology No. 27, Harvard University, Cambridge, Massachusetts. Uribe, Mauricio 2002 Sobre alfareria, cementerios, fases y procesos durante la prehistoria tardia del desierto de Atacama (800-1600

Thompson, Lonnie G., Ellen Mosley-Thompson, Willi Danz gaard, and Pieter M. Grootes 1986 The Little Ice Age as Recorded in the Stratigraphy of theTropical Quelccaya Ice Cap. Science 234:361-364. Thompson, Lonnie G., Mary E. Davis, Ellen Mosley Thompson, Todd Sowers, Keith A. Henderson, Victor S. Zagorodnov, Ping-Nan Lin, Vladimir N. Mikhalenko, R. Kramer Campen, John F. Bolzan, Jihong Cole-Dai, and

Discovery, History, Theory, edited by Robert Arnott, Stan Smith, pp. 223-236. Swets & ley Finger, and C.U.M. Zeitlinger, Lisse, Netherlands. Vivanco, Cirilo 1999 Raqaraqaypata dos poblados de la y Nawpallaqta,

2002 Trepanation inPrehistoric South America: Geographic and Temporal Trends over 2,000 years. In Trepanation:

372LATIN AMERICAN epoca Chanka el el area sur de Ayacucho. InXII Congreso Peruano del Hombre y la Cultura Andina, Vol. 2, edited by Ismael Perez, Walter Aguilar S., andMedardo Purizaga pp. 271-286. Universidad National Ayacucho, Peru.

ANTIQUITY

[Vol.

19, No. 4, 2008

other pukaras are in view, and may have used visual signals to communicate with each other inwartime. 3. huyendo "Pucararo a la maricatatha: phatticatatha, Acogerse fortaleza, o Castillo" (2.275); pucara defined as both "to surrender a fortress" (1.245,

V,

San Cristobal de

Huamanga, Vuille, Mathias, Raymond Bradley, Martin Werner, Richard Healy, and Frank Keimig 2003 Modelling 5180 in Precipitation over the Tropical Americas: 2. Simulation of the Stable Isotope Signal in Andean Ice Cores. Journal of Geophysical Research 108:4175. Wassilowsky, Alexander H. 1999 Investigaciones arqueologicas en la cuenca surdel no In XII (1995-1997). Yanamayo, callejon de Conchucos Congreso Peruano del Hombre y la Cultura Andina, Vol. 2, edited by Ismael Perez, Walter Aguilar S., andMedardo San

maquipatha, 2.275) and "to take over (entrar) a fortress" (1.219); pucara tikhratha, chhichhitha, ccoccotha, and huakhillitha, all mean ing to destroy a pukara (1.245, 2.275). These

terms are

Aymara, not Spanish- or Quechua-influenced neologisms. 4. Across the whole sample, straw from wall mortar yielded slightly later dates than wood charcoal from excava tion units in occupation deposits. At individual sites, however, when both straw and charcoal samples were dated from the same

National Purizaga V., pp. 54-83. Universidad Cristobal de Huamanga, Ayacucho, Peru. Webster, David

site, they did not differ in a consistent fashion; for instance, at Pucarani (N2), charcoal dates were about the same or later than straw dates. This suggests thatwood cura tion is not a major source of error.A possible explanation for the generally late dates from straw is thatwalls were built or

1998 Warfare and Status Rivalry: Lowland Maya and Poly nesian Comparisons. InArchaic States, edited by Gary M. Feinman and JoyceMarcus, pp. 464-470. School ofAmer ican Research Press, Santa Fe, New Mexico. Wells, Lisa E. History of the El Nino Phenomenon as in Flood Sediments of Northern coastal Peru. Geology 18:1134-1137. Wernke, Steven A. 2003 An Archaeo-history ofAndean Community and Land scape: The Late Prehispanic and Early Colonial Colca Val ley,Peru. Unpublished Ph.D. dissertation, Department of Wisconsin. Anthropology, University of Wiessner, Polly, and Akii Tumu 1990 Holocene Recorded 1998 Historical Vines: Enga Networks of Exchange, Rit ual, andWarfare inPapua New Guinea. Smithsonian Insti tutionPress, Washington, D.C. Williams, Patrick Ryan 2001 Cerro Baul: A Wari Center on theTiwanaku Frontier. Latin American Antiquity 12:67-83. 2002 Rethinking Disaster-Induced Collapse in the Demise of the Andean Highland States:Wari and Tiwanaku. World Archaeology 33(3):361-374. Wrangham, Richard, and Dale Peterson 1996 Demonic Males: Apes and the Origins ofHuman Vio lence. Houghton Mifflin, Boston. Zhang, David, Peter Brecke, Harry F. Lee, Yuan-qing He, and Jane Zhang 2007 Global Climate Change, War, and Population Decline inRecent Human History. PNAS 104(49).T9214^19219.

repaired after the initial occupation of pukaras. For example, dates from straw from two different parts of themain outer western wall at Pucarani (N2) are 30 years apart at the very least (1-sigma) and suggest that thiswall was built in at least two separate episodes. Nevertheless, it is possible thatwood

curation may result in artificially early dates for some terraces and occupation strata. Charcoal dates should be seen as rep resenting the earliest probable dates for occupation. This bias serves to emphasize the overall pattern of carbon dates, of consistently late pukara construction and occupation. ascent time was calculated with a GIS 5. Mimimum from ASTER analysis of a digital elevation model (DEM) which calculates 9, using the path distance function, the least cost anisotropic path from a group of source points to any other cell on a raster. Here, source points were defined as cells a minimum of 2 km from the satellite data inArcGIS

with an altitude at or less pukara that lay on the plains?i.e., than 4,000 m and a slope at or less than 5 degrees. The cost of the path was calculated from theDEM inminutes of walk ing time using Tobler's hiker function (Tobler 1975). While the resulting ascent times correlate well with the actual times sidered more it took to ascend each pukara in the field, they should be con in the field we ascended reliable, because

pukaras from different distances, depending on where the modern road lay. 6. Drawing on the estimates of Kolb (1991) and Emerson (1965) for theHawaiian and Mayan areas respectively, I use a rough estimate of 4.5 person-days per cubic meter of wall (including rubble fill). For theworkforce, we can make a very for every two houses on nuclear family used two houses on average, but could only field one able-bodied adult forwall building. The resulting ratios range from 14 days of site: this assumes that a minimal labor per resident wall-builder (at Sinucache, CA2) to about 2,000 (atMallacasi, PKP7), with a median of 245. These esti mates should be considered very rough, since variables such as the quality of wall masonry varied from site to site, but they give an idea of the range of variation. 7. This hypothesis was initially suggested by Rowe (1942) in his analysis of two pukaras near the town of Pucara crude estimate of one wall-builder

Notes
1.A house probably did not equal a household. At the time of Viceroy Toledo's census in 1570, about two centuries after the occupation of pukaras, therewere about 3.5 women, children, and elderly or non-tribute-paying men per tributary adult man in the Colla region (Toledo 1975). This suggests that a nuclear Colla period family of the early Colonial included about 4.5 people. If sex and age distributions were would m similar in the Late Intermediate Period, a minimal household surely have required more than one house structure 3

in diameter. Indeed, houses are most commonly found in groups of two to four on a terrace. 2. Pukaras are preferentially located so that a number of

(Incacancha, PKP5 and Llallagua, PKP1), both of which have defensive walls in unusually poor condition. The idea that late

Arkush]

WAR, CHRONOLOGY,

AND CAUSALITY

INTHE TITICACA

BASIN

373

prehispanic warfare in this region may have included the destruction of pukaras is indirectly supported by Bertonio's Aymara phrases (note 3). ice layer thickness is especially difficult for 8.Measuring earlier years lower in the core, where the ice is compressed and yearly sections are much thinner.Consequently, ice accu mulation values do not correspond well between the two sep arate ice cores from Quelccaya before about A.D. 1300. Oxygen isotope ratios should be resistant to this source of

error, and they do match between the cores, but before about A.D. 1300 an increasing offset between the two curves indi

cates a chronology problem with one or both cores. This problem does not substantially affect the dating of droughts close to the pukara time frame, and hence is not given further discussion here (see Arkush 2005 for a fuller treatment).

Submitted January 20, 2007; Accepted

June 10, 2007.