Journal of Anthropological Archaeology 44 (2016) 56–68
Contents lists available at ScienceDirect

Journal of Anthropological Archaeology

journal homepage: www.elsevier .com/ locate / jaa
Nets and canoes: A network approach to the pre-Hispanic settlement
system in the Upper Delta of the Paraná River (Argentina)
http://dx.doi.org/10.1016/j.jaa.2016.08.003
0278-4165/� 2016 Elsevier Inc. All rights reserved.

Abbreviations: PUD, Paraná Upper Delta; HTN, Hydrographic Transport Network
model; ASN, Archaeological Sites Network model.
⇑ Corresponding author.

E-mail addresses: eapolinaire@gmail.com (E. Apolinaire), laurabastourre@yahoo.
com.ar (L. Bastourre).
Eduardo Apolinaire ⇑, Laura Bastourre
CONICET - División Arqueología, Facultad de Ciencias Naturales y Museo, UNLP, Paseo del Bosque S/N, La Plata City 1900, Buenos Aires Province, Argentina

a r t i c l e i n f o
Article history:
Received 4 November 2015
Revision received 30 April 2016

Keywords:
Network analysis
GIS
Landscape
Upper Delta of the Paraná River
Settlement system
Earthworking
a b s t r a c t

Graph theory-based network analysis provides useful methodological tools for the exploration of several
landscape related issues. In particular, it favours the examination of the topological configuration of
space; that is, the arrangement of its constitutive elements into a relational order. This is an important
aspect of spatiality in terms of its social significance for it reflects, as well as shapes, the way in which
social relations are structured. Based on this approach, a case study from the Paraná River Delta, where
occupation strategies included the construction of earth-mounds and mobility strongly depended upon
water courses, is herein presented. In this scenario, we examined the ways in which spatial arrangement
of settlements and waterways linking them through the landscape reflected and shaped social interac-
tion. More specifically, we evaluated the spatial configuration of the Paraná Delta hydrographic network
in relation to settlement distribution and hierarchy. Thereafter, we found that archaeological sites are
mainly located in highly accessible locations and that the most prominent sites within the settlement
system are located at high centrality areas. Subsequently, we discussed the implications of these results
for the understanding of the emergence of incipient social hierarchies and the significance of earthwork-
ing for the topographic writing of cultural landscapes in the studied area.

� 2016 Elsevier Inc. All rights reserved.
1. Introduction

The deeply entangled relationship between human societies
and their inhabited landscapes has been a major topic of interest
for many archaeologists in the past several decades. On the one
hand, it has been stated that social interactions are woven into net-
works traced over the physical environment in ways in which
socialized landscapes are created (Conkey, 1984; Gamble, 1998,
see also Langley, 2013). The key elements in these socially con-
structed territories are paths and trackways along which informa-
tion flows in order to join individuals and groups together
(Gamble, 1998). On the other hand, it has also been widely recog-
nized that landscape in not merely an external scenario where
social relationships take place but a social production where mean-
ing is imbued into the physical features of the terrain, both natural
and anthropic (Bender, 1993; Cosgrove, 1997; Ingold, 1992, 1997;
Thomas, 2001; Tilley, 1993). Moreover, significant spaces,
landmarks and the pathways that connect them through landscape
topography are attached to meanings and stories which are evoked
over the generations. Such continuum contributes to the construc-
tion and transmission of historical memory and the constitution of
group and individual identity (Bender, 1993; Ingold, 1997;
Thomas, 2001). These socially constituted spaces in turn play an
active and significant role in the organization, reproduction and
transformation of social life. In other words, spatiality at any scale
(landscape, places, settlement systems, architectonic spaces)
enables and shapes certain social relationships, practices and
meanings while other actions and connections are disabled or
ignored (Hillier and Hanson, 1984; Giddens, 1984; Rapoport,
1990, see also Acuto, 2013).

In line with the abovementioned considerations, the configura-
tion of space -that is, the arrangement of its constitutive elements
into a relational order- proves one of the most prominent proper-
ties of spatiality in terms of social significance. Social spaces dis-
play an organization that reflects, as well as shapes, the way in
which social relations are structured (Giddens, 1984). An impor-
tant issue that arises then is how to recognize and depict this topo-
logical configuration. In this regard, graph theory-based network
analyses provide quantitative tools and concepts for analyzing
and representing spatial structurations. An example of such

http://crossmark.crossref.org/dialog/?doi=10.1016/j.jaa.2016.08.003&domain=pdf
http://dx.doi.org/10.1016/j.jaa.2016.08.003
mailto:eapolinaire@gmail.com
mailto:laurabastourre@yahoo.com.ar
mailto:laurabastourre@yahoo.com.ar
http://dx.doi.org/10.1016/j.jaa.2016.08.003
http://www.sciencedirect.com/science/journal/02784165
http://www.elsevier.com/locate/jaa


E. Apolinaire, L. Bastourre / Journal of Anthropological Archaeology 44 (2016) 56–68 57
approach is provided by space syntax, a graph theory-based anal-
ysis drawn upon to assess social aspects expressed in the distribu-
tion and design of architectonic spaces (Bermejo Tirado, 2009;
Dawson, 2002; Hillier, 1996; Hillier and Hanson, 1984). We believe
that graph theory concepts applied to network analysis may also
prove useful tools for the study of social structuration of space in
a broader scale and considering a landscape perspective, a subject
which has been hardly explored but for a few notable exceptions
(Brughmans et al., 2015).

In recent years, network science applications to archaeology
have significantly increased. Such studies place relationships in
the core of our analytical techniques and allow us to approach a
great variety of topics such as hierarchy emergence, settlement
systems and circulation of information, people and goods, among
others (Brughmans, 2013; Collar et al., 2015; Knappett, 2013;
Mizoguchi, 2009). Usually, these relationships are traced among
common features of material culture in order to create relational
webs (Collar et al., 2015). An alternative way of benefiting from
network methods involves shifting the focus of attention to the
spatial properties of the archaeological record and analyzing
inter-site connections in a landscape framework. In this paper,
we explore these ideas via the pondering of a case study from
South American lowlands.

Of late, the traditional point of view regarding South American
lowlands as pristine habitats occupied by egalitarian forager bands
who had no significant impact over the environment has been
challenged by increasing evidence of anthropically modified land-
scapes. Theoretical contributions on Amazonian anthropology,
especially Historical Ecology, have been highly influential in this
regard (Balée and Erikson, 2006; Hornborg, 2005; Hornborg and
Hill, 2011). Pre-Hispanic societies from different regions of South
America developed wetland management strategies involving the
transformation of the landscape through the mobilization of great
volumes of sediments. Such strategies developed into a rich tradi-
tion of earth engineering incorporating mounds, raised fields,
channels and other earthworks spread all along the major South
American basins (Amazon, Orinoco and Paraná-Plata) and tran-
scending cultural and linguistic boundaries (Gianotti and
Bonomo, 2013; Heckenberger and Neves, 2009; Rostain, 2010;
Souza et al., 2016).

Herein, we provide a case study from one of the southernmost
expressions of this earthworking phenomenon, the Paraná River
Delta. In this flood-prone wetland, occupation strategies included
the construction and habitation of earth-mounds and other topo-
graphically elevated areas in an environment where mobility
greatly depended upon water courses. Based on this scenario, we
hereby address how spatial arrangement of settlements and water-
ways linking them through the landscape reflected and shaped
social interaction. For assessing this topologic configuration, we
resort to a graph theory-based network analysis. Spatial structura-
tion of waterways in the study area defined movement and circu-
lation of information, goods and people hence conditioning
interaction possibilities that produce and reproduce social net-
works. In this way, mobility pattern observations provide informa-
tion on social connectivity and, also, inequality (Howey, 2011;
Richards-Rissetto and Landau, 2014). Spaces are usually tied
together in the form of webs within which some locations are more
accessible than others or boast higher potential for controlling
communications. This structuration both influences and reflects
mediation and interpellation abilities of social actors and the range
and nature of the strategies they can implement (Dobres and Robb,
2000; Mizoguchi, 2009). We discuss the implications of these ideas
for the understanding of the emergence of incipient social hierar-
chies and the significance of earthworking in the construction of
cultural landscapes in the study area.
2. Graph theory and network analysis in archaeology

Formal properties of networks can be mathematically
addressed through graph theory. ‘‘Graph” is a term utilized to
describe a twodimensional structure composed of spatially distinct
points or nodes connected by lines or edges. The relevance of this
approach lies on the fact that its use favours the representation of
topological links between network elements beyond the nature or
specific content of these relationships (Cardozo et al., 2009; Hage
and Harary, 1983; Wallis, 2007; Wilson, 2014). Therefore, graphs
have been used to represent structures as diverse as neural circuits,
urban transportation systems, insect colonies or social networks
while currently being widely applied to disciplines such as Physics,
Neuroscience, Sociology, Geography, Computer Science and Eco-
nomics, among others. One of the most remarkable properties of
graphs is centrality. Graph centrality measures are mathematical
methods for quantifying the importance of each node in terms of
its position with respect to the surrounding elements in the net-
work. Two of the most commonly used centrality measures are
betweenness and closeness. These measures indicate how accessi-
ble a location is and the potential for mediation or control that it
may exercise with respect to the traffic between other nodes in
the network (Freeman, 1977; Friedkin, 1991; Sevtsuk and
Mekonnen, 2012).

Formal network analysis through graph theory has been applied
to archaeological research since, at least, the 1960s (Brughmans,
2013). In two influential articles, Pitts (1965, 1979) used measures
of closeness and betweenness to analyze Moscow’s strategic posi-
tion within the river trade network of medieval Russia. The archae-
ological potential of network analysis was clearly recognized in the
1970s by Irwin-Williams (1977), who described the analytical pos-
sibilities offered by network models for quantitative analyses of
prehistoric trade. Particularly, the author explored the potential
of these models when addressing the influence of exchange sys-
tems in prehistoric settlements of northwestern New Mexico. For
his part, Rothman (1987) highlighted the advantages of graph the-
ory in settlement systems analysis in terms of how it allows
researchers to test hypotheses drawn from anthropological theo-
ries bringing into play mathematically objective measures. The
author illustrated this approach by applying graph theory concepts
to the interpretation of regional survey data from south-western
Iran. Influenced by these works, Peregrine (1991) used centrality
measures to explore the evolution of the Cahokia center within
the Mississippi Basin. He visualized this basin as a graph where
lines represent the rivers and nodes correspond to river heads
and junctions. By applying measures of degree, betweenness and
closeness, he argued that the evolution of Cahokia was possible
due to its strategic central position at the confluence of several
major rivers which facilitated control over riverine exchange.

These early applications of graph theory in archaeology devel-
oped discontinuously and had no significant influence on the wide-
spread adoption of network techniques in later archaeological
research (Brughmans, 2013, 2014). Network analyses in archaeol-
ogy faced a significant breakthrough in the last ten years and were
influenced by two research traditions (Brughmans, 2014): social
network analysis (Wasserman and Faust, 1994) and studies of
complex networks in physics (Barabási and Albert, 1999). The
issues addressed in archaeological network analyses are as diverse
as the methods therein applied: spread of information following
the Antonine Itinerary (Graham, 2006), religious innovations in
the Roman Empire (Collar, 2007), impact of natural disasters on
maritime connectivity in the Aegean Bronze Age (Knappett et al.,
2008), identification of social and cultural boundaries in Papua
New Guinea (Terrell, 2010), distribution of Roman pottery
(Brughmans and Poblome, 2016), social interactions between Near



58 E. Apolinaire, L. Bastourre / Journal of Anthropological Archaeology 44 (2016) 56–68
East sites during Epipalaeolithic and early Neolithic times (Coward,
2013), centralization of lowland Maya political networks (Munson
and Macri, 2009), procurement and distribution of pre-Hispanic
Mesoamerican obsidian (Golitko and Feinman, 2015) and intervis-
ibility of Iron Age and Roman sites in southern Spain (Brughmans
et al., 2015), among others.

Beyond thematic diversity, several specific contributions
address issues of particular relevance a propos of the aims of the
study hereby conducted: transportation systems and mobility
throughout the landscape and emergence of social hierarchies.
Similarly to Peregrine (1991), Jenkins (2001) resorted to measures
of degree, closeness and betweenness to explore the properties of
the Inca transport network. Particularly, the author discussed the
topological advantage of administrative centers, production sites
and storage sites placed in strategic locations within the Inca road
system. Similarly, Isaksen (2007, 2008) analyzed the Roman road
system in southern Spain and highlighted the potential of GIS
when studying this sort of geographical networks (Brughmans,
2014; Knappett, 2013). For his part, Mizoguchi (2009) applied mul-
tiple centrality measures to assess the emergence of centralized
hierarchies in the Japanese initial Kofun period. Such contribution
concludes by determining that the topological position of political
units within a social network can be a significant cause of the
emergence of hierarchical relationships among them. Centrality
measures are the most commonly used methodological tools for
studies such as these, where the basis consist of the identification
of nodes that feature either a better access to information/
resources or an intermediate position that allows them to control
the connections between other nodes and, hence, the circulation
within the network.

Stemming from a different tradition, another interesting contri-
bution to archaeology involving graph theory is space syntax. This
approach studies the ways in which built spaces constituting archi-
tectural ensembles are related and organized. The point of such
pondering is to assess the underlying social logic of such structura-
tion (Bermejo Tirado, 2009). In this framework, the relationship
between society and space is a dynamic one, where each variable
modifies and restructures the other (Bafna, 2003). Originally devel-
oped by architects (Hillier, 1996; Hillier and Hanson, 1984), space
syntax has been used in archaeology to examine spatial configura-
tion of past public and domestic architecture (Dawson, 2002;
Bermejo Tirado, 2009).

Finally, a graph theory derived tool, circuit theory, has been
recently adapted from electrical engineering to biology sciences
and, thereafter, to archaeology. This approach is used in biology
to model environment connectivity and predict patterns of move-
ment, gene flow and genetic differentiation among plant and ani-
mal populations (McRae, 2006; McRae and Beier, 2007). In
landscape archaeology, circuit theory has been drawn upon as a
novel and complementary approach to least-cost models in order
to study mobility patterns and trace past pathways through land-
scape (Howey, 2011; Thayn et al., 2016).
3. Waterways, graphs and centrality in the Paraná Delta

The Paraná River is, with a total length of more than 3500 km
and an extensive delta system located in its lower section, one of
the most important fluvial courses of South America. The Paraná
Delta consists of a dense and complex fluvial network comprising
rivers, streams, ponds and temporary narrow water courses sur-
rounding elevated landforms (Iriondo, 2004). In this environmental
context, such network is, both at present and in the past, the main
transport route for human mobility. Sixteenth-century ethnohis-
torical sources describe canoe mobility among the Chaná, the
Timbú, the ChanáBeguá, the ChanáTimbú and other groups that
inhabited the region (Fernández de Oviedo y Valdez, 1547;
Medina, 1908; Ramírez [1528], 2007; Santa Cruz [1540], 1918;
Schmidl [1567], 1980). Dugout canoes found in the Paraná Delta
and Río de la Plata River are material records evidencing these
practices in the pre-Hispanic past (Bonomo, 2012; Serrano,
1950). Furthermore, navigation is the main form of transportation
among the ‘‘islanders” inhabiting the region even at present.

Archaeological research in the Paraná Upper Delta (hereafter
referred to as PUD) enabled the detection of more than 60 sites
typically emplaced at high landforms of different origin (mainly
mounds, locally known as ‘‘cerritos” and levees) and often located
near the intersection of two or more fluvial channels. Such
emplacement would allow easy access to waterways (Bonomo,
2012; Gianotti and Bonomo, 2013). In addition to exhibiting func-
tional and seasonal differences, the sites presumably reflect a set-
tlement hierarchy in which some localities are characterized by the
size or quantity of the mounds therein emplaced, the effort
invested in their construction, their temporal stability, the ritual
use of burial areas and the presence of exotic goods. Based on this
evidence, the fact that indigenous groups who inhabited PUD (eth-
nohistorically known by the generic term of ChanáTimbú) and
their predecessors centuries prior to the Hispanic conquest (related
to the Goya-Malabrigo archaeological entity) lived in the frame of a
sociopolitical organization corresponding to a ranked society has
been proposed. In such societies, high ranked individuals exert reg-
ular and repetitive authority but their followers are not bound to
obey them (Chapman, 2003). In addition to the archaeological evi-
dence, this proposal is based on several allusions, found in
sixteenth-century chronicles, to the existence of charismatic and
powerful leaders with the ability to form alliances with each other
and coordinate regional exchange networks (Bonomo et al., 2011;
Politis and Bonomo, 2012).

Due to the importance of fluvial transport in the study area, we
believe that a thorough analysis of the Paraná fluvial network is
crucial for the exploration of human settlement, mobility and
exchange and, hence, for understanding the way in which social
interactions were structured over the landscape. Therefore, we
evaluate the spatial configuration of the waterway transport sys-
tem in relation to distribution and hierarchy of the settlements
and circulation of foreign materials. We pursue these goals by
way of resorting to an approach that combines social network
analysis concepts derived from graph theory with GIS tools. This
is accomplished by making use of software originally developed
for urban network analysis. With such software, it is possible to
model both topological and metric (geographical) networks which
therefore favours the combination of measurements derived from
network analysis and the analytical advantages of quantitative
spatial variables.

In this paper, we model the Paraná Upper Delta fluvial network
as a graph in which nodes represent waterway intersections and
edges stand for the fluvial streams themselves. Based on this
model, which we call Hydrographic Transport Network, an Archae-
ological Sites Network model is developed to explore the relation-
ships between archaeological sites (nodes) connected by fluvial
channels (edges). Two centrality measures, betweenness and
closeness, are calculated via these models in order to account for
differences in accessibility and potential for controlling communi-
cations over the landscape. Considering the fact that spatial struc-
turation of settlements and mobility reflected as well as shaped
social interactions, several expectations can be outlined. Firstly, it
is hypothesized that archaeological site distribution will relate to
variations in the degree of accessibility within the hydrographical
network. Secondly, the most prominent sites within the settlement
system are expected to be located in areas with high communica-
tion potential that may also allow better control over the flow of
information, goods and people. Therefore, it is also expected that



E. Apolinaire, L. Bastourre / Journal of Anthropological Archaeology 44 (2016) 56–68 59
central points of the landscape concentrate the highest amount of
exotic goods since these points are better connected and present
an advantageous position within the exchange network.

4. Environmental framework

The Upper Delta of the Paraná River covers the area between
Diamante City (32�41010.9600S and 60�38016.3400W) to the north
and the Paraná Pavón River to the south while it is bounded to
the east and west by the cliffs that define the transition to the high
plains of Entre Ríos and Santa Fe Provinces (Fig. 1). The study area
is characterized as a gently sloping floodplain with a complex
hydrological regime subject to periodical floodings. Extensive low-
lands represented by permanent and temporary ponds, floodplains
and ancient tidal plains which surround elevated landforms such
as fluvial banks, leeves, meander scrolls and sand dunes can be rec-
ognized (Castiñeira et al., 2014; Cavallotto et al., 2005).

The PUD is part of the littoral complex that developed in the
lower section of the Paraná River over four phases (Iriondo,
2004). The fluvio-eolian initial phase is represented today by the
Pleistocene Flood Plain, the oldest part of this complex. Subse-
quently, the middle Holocene marine ingression led to the devel-
opment of a sand barrier, a lagoon, minor tributary deltas and
estuaries and well developed regressive deposits. During the suc-
cessive estuarine phase, an extensive tidal plain developed in the
center of the littoral complex. Finally, the current fluvial-deltaic
period is marked by channel and deltaic deposits advancing into
the Río de La Plata. Within the study area, located to the north of
this littoral complex, several geomorphological units correspond-
ing to this evolutionary sequence can be recognized (Iriondo,
2004) (Fig. 1). The Pleistocene Flood Plain is the result of a
long evolution under flood dynamics. It consists of a vast
Fig. 1. Location of the Upper Paraná Delta within the southern cone of America a
hindered-drainage plain characterized by a large number of shal-
low lakes, ponds and swamps cut by minor adventitious channels.
These channels are bordered by well developed levees which iso-
late them, except during floods events, from active channels. South
of this unit and during the retreat of the sea, a set of beach ridges
consisting of a sequence of beaches juxtaposed or separated by
mudflats was formed. Adjacent to these beaches, a tidal plain com-
posed of estuarine sediments deposited after the higher phase of
the Holocene ingression was developed. Such tidal plain is charac-
terized by subparallel tidal channels presenting straight segments
and well-defined angles. This particular unit is currently subjected
to fluvial dynamics with floods that partly run through the tidal
channel network and modify it in some areas. Throughout the late
Holocene and at present, the fluvial dynamics of the Paraná River
and its tributaries have resulted in bar and meander belts as well
as a plain of narrow meanders. The first unit developed in high
energy environments during the current fluvial-deltaic phase and
is the most active geomorphological unit of the study area. It is
formed by bedload deposition along the main channels of the
Paraná River. This deposition generates bars that eventually
become permanent islands due to vertical aggradation. The plain
of narrow meanders, characterized by a large number of sinuous
low-energy channels following flow-paths that operate indepen-
dently over considerable distances, represents the oldest portion
of the floodplain and was generated by the lateral migration of
quite active distributaries of the left bank of the main channel. At
present, this process of lateral migration continues to develop
reworking sediments of older units, principally the Pleistocene
Flood Plain and the Tidal Plain. However, as evidenced by persist-
ing features of older units, this alteration of forms is not intense
(Iriondo, 2004; Iriondo and Kröhling, 2008; Milana and Kröhling,
2015; Paira and Drago, 2007).
nd geomorphological units of the study area (modified from Iriondo, 2004).



60 E. Apolinaire, L. Bastourre / Journal of Anthropological Archaeology 44 (2016) 56–68
5. Archaeological framework

As a result of the research project conducted by Drs. Politis and
Bonomo in the Diamante, Victoria, Gualeguay (Entre Ríos Province)
and San Jerónimo (Santa Fe Province) Departments, 84 archaeolog-
ical sites were detected of which most are located in the insular
area of this geographical region. Radiocarbon dates obtained so
far yield a time span of ca. 1920-500 14C years BP which reveals
that human occupation probably took place after the establish-
ment of the current geographical configuration (Apolinaire et al.,
2016; Bonomo, 2012; Bonomo et al., 2010, 2014; Castiñeira et al.,
2012, 2014; Gianotti and Bonomo, 2013; Politis et al., 2011). There
is clear evidence of earthmounds, which represent about 50% of the
detected sites, since, at least, ca. 1900 BP (Apolinaire et al., 2015;
Castiñeira et al., 2014). Such mound-building activity has been
associated with Goya-Malabrigo, an archaeological entity probably
corresponding to the ancestors of historical ethnic groups generi-
cally known as Chaná-Timbú. Goya-Malabrigo has been character-
ized by a particular pottery style marked by emblematic forms
such as ‘‘campanas” and zoomorphic appendices (Ceruti, 2003;
Politis and Bonomo, 2012; Serrano, 1950). Additionally, a mixed
economymainly based on fishing and semi-aquatic mammal hunt-
ing (Bastourre, 2015) complemented by gathering and maize, bean
and squash horticulture (Sánchez et al., 2013) has been suggested.
This aquatic oriented economy is associated to the development of
fluvial technology consisting of fishing nets, harpoons and dugout
canoes.

Archaeological research at PUD enabled the identification of a
settlement system where sites presenting different functions, hier-
archy and occupation intensity can be recognized (Politis and
Bonomo, 2012). Major residential settlements were located at the
highest landscape positions (unaffected by floods except in times
of extraordinary inundations) and consist of anthropogenic
mounds, as well as natural landforms such as dunes and levees.
These settlements, interpreted as actual villages, were semi-
permanently occupied during prolonged periods of stability fol-
lowed by abandonment and subsequent reoccupation. Second
order settlements consist of less intensive occupations in naturally
elevated landforms slightly lifted by human activity. These sites
were not periodically reoccupied and their function could be
related to the exploitation of certain resources. Also, there were
other settlements, perhaps oriented to specific activities, consis-
tent with more sporadic, shorter and less dense occupations than
those characteristic of the previous case (Bonomo et al., 2011;
Politis and Bonomo, 2012). Archaeological topographic surveys at
PUD enabled the design of digital elevation models of 20 mound-
like structures and the calculation of their size and volume
(Bonomo et al., 2011; Castiñeira et al., 2014). These structures
are either isolated in the landscape or aggregated in groups of
two or three in which one is usually larger than the others. These
mound-like structures typically exhibit elliptical shapes and,
although some have been altered by erosion, their heights range
from 0.5 to 2.2 m. Their sizes and volumes present considerable
variations marked by volume values ranging from 140 to
3912 m3 and major axis lengths between 35.4 and 84.2 m.

Settlements within the study area have not been isolated but
integrated into regional interaction systems and connected to
supra-regional exchange networks. This is reflected in the presence
of foreign rocks with natural outcrops located medium to long dis-
tances away from the archaeological sites: sandstones from the
Ituzaingó Formation to the north (middle Paraná River); silicified
limestones, sandstones and basalts that outcrop along the Uruguay
River to the east; quartz from the Sierras de Córdoba to the west
and orthoquartzites from the Balcarce Formation to the south
(Bonomo and Blasi, 2010). In fact, considering the absence of nat-
ural outcrops in the study area, lithic raw materials are rare at
archaeological sites and mainly have a foreign origin. The informa-
tion available for Goya-Malabrigo from both PUD and other sectors
within its dispersal area indicates that, in addition to rocks, other
items (such as copper artifacts, malachite beads, textiles and
domestic camelids) have circulated among these supra-regional
networks, thus evidencing links with the southern Andean region
and Central Sierras of Argentina (Bonomo et al., 2011).

Politis and Bonomo (2012) suggested that social groups associ-
ated to the GoyaMalabrigo archaeological entity lived in the frame
of a socio-political organization corresponding to what is usually
known as ranked society (sensu Chapman, 2003). In these societies,
highly ranked individuals enjoy differential access to exotic prod-
ucts and prestige goods yet all individuals equally access the same
basic resources regardless of their social position. Moreover, chiefs
have the role of preventing fragmentation within the group, coor-
dinating collective work and controlling long-distance exchanges.
Such hypothesis is based on archaeological evidence of burials con-
taining prestige goods, the existence of a hierarchical settlement
system that could reflect social distinctions and the presence of rel-
atively large and grouped mounds for the construction of which
the organization of communal work must have been required. Fur-
thermore, sixteenthcentury ethnohistorical documents mention
the existence of chiefs or ‘‘mayorales” -identified by their power,
representativeness and paraphernalia- who had the ability to fed-
erate several groups to engage in warfare against the Spanish con-
querors (Bonomo et al., 2011; Fernández de Oviedo y Valdez
[1547], 1851; Lopes de Sousa [1530-32] 1839; Medina, 1908;
Schmidl [1567], 1980).
6. Material and methods

The hydrographic network of the Paraná Upper Delta was mod-
elled using GIS tools. To this aim, we resorted to information from
the Instituto Geográfico Nacional (IGN) (SIG250) as a source of
hydrographic vector data. This fluvial network was extended and
detailed by means of the digitization of minor streams. For such
purpose, satellite images obtained from Google Earth (Google
Inc.) and Bing Maps (Microsoft Corp.) servers were worked at a
scale of 1:75,000. These images display dates ranging from 2012
to 2013 and were obtained from months during which the Paraná
River level at Rosario City Port ranged between 2.25 and 3.25 m.
The obtained vector data were processed with Network Analyst
package (ArcGIS, ESRI Inc.) to attain a Hydrographic Transport Net-
work model (hereafter HTN) in which lines represent waterways
and nodes herald places where the waterways intersect (junctions)
or stream endpoints can be found. In the HTN, lines are undirected
given the fact that they adopt no fixed direction of flow and con-
nect nodes in both directions. In this model, flow impedance is
defined by waterway length measured in meters. The lagoons were
modelled as central nodes connecting all the lines that converge
into them.

Centrality measures for nodes in the HTN were calculated using
the Urban Network Analyst package (Sevtsuk and Mekonnen,
2012). Two measures were considered: closeness centrality and
betweenness centrality. Closeness centrality measures the proxim-
ity of a node to all other nodes in the network. Formally, it can be
defined as:

Cr
c½i� ¼

1P
j2G�fig;d½i;j�6rd½i; j�

Closeness centrality Cr
c½i� of a node i in a graph G is the inverse of

the sum of the distances between i and all other nodes that are
reachable in G within radius r following the shortest paths
(Sevtsuk and Mekonnen, 2012). For this study, the distance was



E. Apolinaire, L. Bastourre / Journal of Anthropological Archaeology 44 (2016) 56–68 61
measured in meters and the radius considered included the entire
network. Closeness centrality quantifies the ease with which a
node can reach and be reached by every other node in the network.
Specifically, it can be regarded as a useful measure for assessing
the importance of a given node in terms of its accessibility.

Betweenness centrality (Freeman, 1977), Cr
I½i�, indicates the fre-

quency with which a node lies on the shortest path between any
pair of other nodes in the network. It is defined as:

Cr
I½i� ¼

X

j;k2G�fig;d½j;k�6r

njk½i�
njk

where njk is the number of shortest paths from node j to k in G and
njk½i� is the number of shortest paths from j to k passing through i,
with j and k lying within the radius r from i (Sevtsuk and
Mekonnen, 2012). Unlike closeness, which illustrates the ease with
which a node can reach and be reached by other nodes, between-
ness indicates the potential of a node to be ‘‘on the way” between
any other pair of nodes in the network, when the shortest paths
are considered. Elevated betweenness values for a given node sug-
gest a high degree of control of such node over the network since
many other nodes depend on it to connect with each other. In other
words, a node with high betweenness centrality mediates interac-
tions within the network to a greater extent than others. In a trans-
port network, these key nodes may have the potential to influence
the flow of goods, information and people (Isaksen, 2008;
Mizoguchi, 2009).

A raster image with a cell size of 20 m was generated from the
interpolation (kriging method) of the centrality values calculated
for the HTN nodes. Thus, centrality values could be assigned to
each cell of the matrix, obtaining closeness and betweenness maps
with values for each point of the study area. Finally, centrality val-
ues were extrapolated to archaeological sites based on their geo-
graphic position within these maps. We assume that these values
indicate the accessibility of a settlement and its ability to control
the hydrographic transport network. Since HTN is a geographic
model independent from archaeological variables, site centrality
values obtained from closeness and betweenness maps are not
affected by any archaeological sampling biases.

Based on the HTN, a second model was developed to account for
connections among archaeological sites. In this case, each site was
represented as a node linked to the HTN edges through the shortest
line; that is, to the closest point along the nearest water course. On
this basis, the Archaeological Sites Network model (hereafter ASN)
was obtained. In the ASN model nodes represent sites and lines,
plotted on the hydrographic network, correspond to the shortest
waterways connecting the nodes together. Finally, centrality mea-
sures for nodes in the ASN were computed. Unlike the HTN model,
where closeness and betweenness measures indicate accessibility
and control over the fluvial landscape, in the ASN model, these
measures reflect site centrality in relation to other archaeological
sites in the study area. More precisely, closeness indicates the ease
with which a site can access or be accessed by another one while
betweenness reveals its potential for controlling communications
among sites.

Thereafter, we relate the results obtained from these analyses to
several features of the regional archaeological record that are of
relevance for the aims of this work: distribution of archaeological
sites, settlement hierarchy and presence of foreign materials. We
considered a total of 58 archaeological sites detected during sur-
veys in PUD and described in several publications (Bonomo et al.,
2010, 2014; Gianotti and Bonomo, 2013). Although the southeast
sector was not as deeply prospected, these surveys were intensive
in most of the study area (Bonomo et al., 2010). Due to the regional
scale utilized, localities with archaeological sites separated by
distances of less than 1 km were considered a single entity thus
reducing the sample considered to 51 analysis units.

As regards site distribution, we evaluated through spatial auto-
correlation (Global Moran’s I) whether its spatial pattern is ran-
dom, clustered or dispersed, using centrality values of sites
(obtained from the HTN model) as associated attributes. In other
words, we examine if archaeological sites are more clustered or
dispersed than would be expected if underlying spatial processes
were random. In addition, we assessed whether archaeological
localities tend to be emplaced in geographical areas marked by
high closeness and/or betweenness scores. A Z-test was performed
to examine whether closeness values for archaeological sites
(obtained on the basis of the HTN) correspond to a random sample
drawn from the set of points that make up the study area (namely,
the set of cell values comprising the closeness raster map), or
instead they are significantly higher than the mean value for the
raster. In the case of betweenness, whose distribution is not nor-
mal, a Mann-Whitney test was performed to the same end. To
avoid potential biases associated to the archaeological survey,
these statistical tests were calculated excluding the least pros-
pected sectors from centrality maps.

To examine settlement hierarchy, we considered mound size as
an indicator of the importance of each site within the settlement
system. For such purposes, we took the area and volume values
published for 20 mounds (corresponding to 15 analysis units) of
the study area (Bonomo et al., 2011) into account. The mound area,
calculated as an ellipse using the length of the major and minor
axes of the structure, is herein assumed to represent the inhabit-
able and anthropogenically modified surface. The mound volume
was deemed to be an indicator of the effort invested in earth move-
ment for the construction of the mound (Bonomo, 2012). In local-
ities presenting more than one mound, the areas were summed
and the volume was calculated as the average volume multiplied
by the number of mounds. Then, area and volume values for these
sites were correlated to their centrality scores, computed for
both the HTN and the ASN models. Pearson and Spearman’s
correlation coefficients were used for closeness and betweenness,
respectively.

Finally, we examined the relationship between the centrality
scores for each site and the presence of foreign rocks. Since not
all archaeological localities were studied thoroughly, there are
not enough data available to compare frequency and diversity of
lithic raw materials between sites. Therefore, the sample of sites
was divided based only on the presence or absence of foreign rocks.
Then, the centrality mean values of these subsets, calculated for
both HTN and ASN models, were compared in search of any signif-
icant differences between them. For closeness scores comparisons,
a T-test was performed while betweenness was analyzed by means
of the Mann-Whitney test.
7. Results

The distribution of archaeological sites within the study area is
related to the topological structure of the hydrographic network.
Fig. 2 illustrates the HTN model and the gradient values of close-
ness (a) and betweenness (b) calculated through interpolation. As
regards closeness, spatial autocorrelation results (Moran’s
I = 0.61, z = 6.06, P < 0.01) indicate that archaeological sites are
not randomly distributed but display a clustered pattern. That is,
sites tend to be spatially aggregated according to their closeness
values more than would be expected by random chance. As shown
in Fig. 2a, archaeological sites are mainly located in areas featuring
high closeness scores. In fact, Z-test (zP 3.24, P < 0.01) indicate
that closeness mean value obtained for site dataset is significantly



Fig. 2. Distribution of archaeological units in the study area (n = 51) and interpolated centrality values for the Hydrographic Transport Network model.

62 E. Apolinaire, L. Bastourre / Journal of Anthropological Archaeology 44 (2016) 56–68
higher than population mean calculated for the set of HTN raster
cells. Thus, the distribution of settlements seems to be related to
their potential access to other locations in the landscape. In the
case of betweenness centrality, statistical results indicate that this
variable is not related to the spatial distribution of archaeological
sites. In addition, sites are not necessarily located in areas of high
betweenness.

In Table 1, we present closeness and betweenness values for the
archaeological sites, calculated for both models (HTN and ASN),
along with the areas and volumes of the mounds and the pres-
ence/absence of foreign rocks (Bonomo et al., 2011). Closeness val-
ues of the sites in the HTN are strongly correlated with both the
area (r = 0.726, p < 0.01) and the volume of the mounds
(r = 0.649, p < 0.01) (Fig. 3a and b). That is, sites with larger
mounds tend to be located at the most accessible places of the
study area. When considering the archaeological sites network
(ASN), we observed a weaker but significant positive correlation
between closeness and volume (r = 0.470, p < 0.05) as well as
betweenness and area (r = 0.479, p < 0.05) (Fig. 3 c and d). In other
words, mound size may also be related to its potential to reach and
be reached by other sites in the network and to mediate the inter-
actions among them. In no case was there a significant correlation
between mound size and betweenness values calculated for the
HTN.

When considering the ASN (Fig. 4), we noted that closeness and
betweenness scores for the set of archaeological sites featuring for-
eign materials are significantly higher than the values obtained for
cases in which these materials were not detected (t = 2.033,
p < 0,05; u = 395, p < 0.1). However, in the case of HTN, centrality
scores present no significant differences between both sets of
sites.
8. Discussion

In this work, the Paraná Upper Delta hydrographic network was
modelled as a graph using GIS tools. Such network is assumed to
have served as a transportation system that facilitated regional cir-
culation and exchange of goods, information and people between
roughly synchronous settlements. This assumption involves two
problems:

(1) It uses current data to work out a model for the ancient
hydrographic network without considering possible topo-
logical configuration changes that may occur in a dynamic
fluvial system. Paleoenvironmental reconstructions indicate
that the Paraná Upper Delta begins to outline its current
configuration circa 4000 14C years BP; that is, after the last
Holocene marine transgression. According to available
radiocarbon dates and as evidenced by the archaeological
remains recovered from fluvial flood sediments (Castiñeira
et al., 2014), human occupation occurred after such geo-
graphical configuration was established. Although river net-
work structures change through time, we believe that, in the
time span considered, these alterations were not as substan-
tial as to modify the overall network configuration. In fact,
much of the study area corresponds to geomorphological
units currently subjected to low energy fluvial action by
which the alteration of inherited landforms is not intense
(i.e. Pleistocene flood plain, beach ridges, tidal plain and
plain of narrow meanders). In some cases, the fact that the
current hydrographic network flows through the ancient
channels that developed during the post-transgressive estu-
arine phase can even be observed. However, the bar and



Table 1
Data considered for each analysis unit.

Analysis units HTN ASN Area, m2 Volume, m3 Foreign raw materials

Closeness � 10,0000 Betweeness Closeness x 100,000 Betweeness

Cerro de Diego 0.016366 6491.378 0.024237 44 743.968 270.800 Absence
Cerro Lote 11 0.018678 14100.441 0.023073 10 1017.593 325.000 Absence
Cerro El Lucerito 0.018787 5522.407 0.035137 0 1371.588 583.800 Absence
Cerro Justo Norte 0.017887 16328.697 0.029818 12 1570.074 897.300 Absence
Cerro Los Cardos 0.016520 25416.548 0.024435 152 1270.272 901.000 Presence
Cerro Grande 0.019757 31334.808 0.028012 214 1756.669 927.900 Absence
El Cerrito de Puerto Esquina 0.018000 13244.353 0.024104 212 2672.490 973.100 Presence
Cerro de las Cañas 0.016358 5998.290 0.024113 36 1636.463 1232.500 Absence
Cerro Grande de la Isla de los Marinos 0.021189 8631.055 0.032149 30 2158.376 2088.200 Absence
Cerro El Durazno 1 0.021563 1063.572 0.029872 33 2296.135 2660.300 Presence
Cerro Puesto Acosta 0.022793 13594.128 0.030988 99 3651.308 2978.100 Absence
Cerro El Castaño 0.021748 17201.869 0.033006 284 4219.913 3020.200 Presence
Cerro Barrancas 0.017337 6255.802 0.032402 0 1947.065 3064.800 Absence
Los Tres Cerros 0.022256 7859.110 0.025072 106 6312.434 3798.300 Presence
Cerro Tejeira 0.019661 9841.714 0.033833 184 3290.567 3912.600 Presence
A El Espinillo 1 0.018449 4163.601 0.029834 8 – – Presence
A El Espinillo 2 0.018488 5140.454 0.029981 42 – – Absence
A El Espinillo 3 0.018990 4790.719 0.030616 121 – – Absence
Arroyo Las Tejas 0.017644 8217.363 0.021970 102 – – Absence
Boca de la Sangría 0.021904 3272.725 0.031278 0 – – Absence
Cerro Arena 0.018752 2348.725 0.025684 90 – – Absence
Cerro Bella Vista 0.019051 3849.607 0.030218 0 – – Absence
Cerro Camino 0.018904 2779.594 0.025477 32 – – Absence
Cerro Chico 0.019699 34817.078 0.035022 226 – – Presence
Cerro Cortada Pelegrini 0.017882 1651.073 0.022777 0 – – Absence
Cerro de Arena 0.020355 13405.304 0.033347 258 – – Absence
Cerro de Vázquez 0.018462 13257.694 0.023690 80 – – Absence
Cerro de Zamora 0.015383 13944.768 0.020823 0 – – Absence
Cerro el Durazno 2 0.020370 4906.858 0.031601 82 – – Absence
Cerro El Manolo 0.020291 7766.756 0.027359 152 – – Absence
Cerro Farall 0.013273 1221.801 0.019176 0 – – Absence
Cerro Rodríguez 0.018977 4890.693 0.030460 0 – – Presence
Cerro Tapera Vazquez 0.013192 1462.587 0.019358 0 – – Presence
Cerro Las Moras 0.021602 5821.994 0.032334 160 – – Absence
El Refugio 0.023936 48507.367 0.034513 403 – – Absence
La Banqueta 0.021311 28022.498 0.034837 116 – – Absence
La Ciega 0.014000 0.000 0.016334 0 – – Absence
La Gotera 0.020470 24927.311 0.034362 174 – – Absence
La Horqueta 0.016809 4883.964 0.024896 54 – – Presence
La Tortuga 1 0.021353 4495.965 0.031234 56 – – Absence
La Tortuga 2 0.021488 5456.028 0.030674 0 – – Absence
La Tucura 0.017299 15797.717 0.022464 200 – – Absence
Laguna de los Gansos 0.017025 5851.275 0.028143 8 – – Presence
Laguna Grande 0.020479 19632.172 0.033910 232 – – Presence
Los Baños 0.021758 24888.174 0.032944 0 – – Absence
Los Dos Cerros 0.017995 8799.604 0.029367 88 – – Presence
Los Laureles 0.019965 15800.887 0.033564 178 – – Presence
Los Remanses 0.017293 13569.963 0.023810 184 – – Absence
Puesto Gómez 0.022088 18003.418 0.032928 260 – – Absence
Puesto La Camiseta 0.016485 6144.892 0.022640 0 – – Absence
Rincón Saldana 0.015632 14466.500 0.021127 98 – – Absence

E. Apolinaire, L. Bastourre / Journal of Anthropological Archaeology 44 (2016) 56–68 63
meander belts unit, comprising a narrow strip along the
main channels of the Paraná, is permanently modified by
intense fluvial action (lriondo, 2004). To consider the mod-
ern anthropic alterations of the fluvial system, such as
bridges (particularly the Rosario-Victoria Bridge), local
embankments and hydroelectric dams located at the upper
Paraná River, is also necessary. In sum, even though the fact
UPD hydrographic configuration has not been substantially
modified since human occupation seems probable, it should
be noted that the model proposed herein loses elucidatory
potential as it is applied to increasingly early dates.

(2) It considers that archaeological sites represent more or less
synchronous occupations within the same settlement sys-
tem. Although the time scale used in this study covers a per-
iod of at least 1400 years (1900 to 500 14C years BP), it is
noteworthy that many sites were repeatedly occupied for
extended time intervals. This is the case of Los Tres Cerros
1 marked by 21 radiocarbon dates ranging between 1227
and 560 14C years BP which evidence an occupational history
of over 650 years (Scabuzzo et al., 2015). Additionally, most
of the sites considered herein were assigned to the same
archaeological entity (Politis and Bonomo, 2012), thus sup-
porting the assumption that their settlements were linked
together.

Under these assumptions, the position of the archaeological
sites within the hydrographic network model was analyzed. Such
assessment resulted in finding that the distribution of sites is
clearly related to the network topology: sites are preferably located
at places that allow easy access to other points of the landscape,
considering distance as the main impedance factor. Therefore, we
argue that transportation network structure was an important fac-
tor in decisions related to settlement location. Thus, central loca-
tions were probably favoured for allowing both a better access to



Fig. 3. Scatter plots showing the relationship between mound size and centrality values.

64 E. Apolinaire, L. Bastourre / Journal of Anthropological Archaeology 44 (2016) 56–68
the surrounding resources and better communications with other
settlements connected through the transport network.

These results can be discussed in light of the features that have
been described for the Goya-Malabrigo archaeological entity and
which are shared with Arawak societies extended throughout
much of the South American lowlands (Politis and Bonomo,
2012). Without entering into the discussion of GoyaMalabrigo ori-
gins and its possible Arawak affiliation, we believe that, from a
comparative approach, the ethnographic, historical and archaeo-
logical information corpus on Arawak societies (Eriksen, 2011;
Hill and Santos-Granero, 2002 Hornborg, 2005; Neves, 2001) is a
useful interpretative framework for further exploring our results.
In the first place, the location of settlements in accessible and
well-connected areas within the fluvial transport network is con-
sistent with the depiction of a society characterized by its connect-
edness, openness and expansiveness along waterways. This is
usually expressed by the development of broad alliances among
local and regional groups (both at intra- and inter-ethnic level)
and the widespread occurrence of regional and even interregional
social formations (Santos-Granero, 2002). The establishment of
alliances with neighbouring groups along the rivers is possibly
what generated extensive interaction networks between commu-
nities linked by kinship, trade and other mechanisms (Eriksen,
2011; Hornborg and Hill, 2011). Another characteristic of the
so-called Arawak ‘‘ethos”, and linked to the abovementioned
attributes, is the suppression of endo-warfare phenomena. These
features are absent among other native communities of the South
American lowlands in which social bonds tend to fragment thus
resulting in political systems failing to coalesce into larger social
formations and where institutionalized cycles of vendettas, orga-
nized raiding and other forms of collective violence are part of
the constitution of social identities (Santos-Granero, 2002). Some
features of this Arawak ‘‘ethos” can be distinguished in the
GoyaMalabrigo archaeological entity and the historically related
ChanáTimbú groups: the existence of exchange networks at local,
regional and supraregional levels evidenced by the circulation of
foreign materials; the absence of internal conflicts among different
ChanáTimbú groups, even though violence episodes did transpired
against Guaraníes, Charrúas and Querandíes (as shown by
sixteenth-century documents); and the ability to generate regional
confederations to raid Spanish settlements (Bonomo and Blasi,
2010; Bonomo et al., 2011; Politis and Bonomo, 2012). In this arti-
cle, we propose that the existence of these interaction networks
was facilitated by the location of settlements in central parts of
the hydrographic network.

Within this hydrographic system, certain mound sites stand out
due to their size. This earthworks implied a greater mobilization of
resources (in terms of earth movement and human labour) to erect
enduring structures in the landscape probably occupied by a
greater number of people in a relatively stable fashion for several
generations. In some cases, such mounds have also been chosen
as burial areas, signifying these places in a special way. Through
network analysis, it was possible to observe that, furthermore,
the topological position of these sites is distinguished by its cen-
trality within the hydrographic network. That is, not only do the
settlements tend to be emplaced in accessible areas but, also, the



Fig. 4. Archaeological Sites Network model and archaeological sites distinguished by betweenness values and presence of foreign raw materials.

E. Apolinaire, L. Bastourre / Journal of Anthropological Archaeology 44 (2016) 56–68 65
places of greater accessibility appear to be chosen for the establish-
ment of the greatest and most enduring and meaningful mounds.

We believe that the concept of ‘‘topographic writing”
(Santos-Granero, 1998) becomes of interest when discussing these
results. The author stresses the importance, among Arawak groups,
of this form of imbuing cultural landscapes with historical
knowledge. The topographic writing is based on the existence of
landscape features (so-called topograms and topographs) that act
as mnemonic devices to evoke and transmit historical events and
processes, especially those in which spatial dimension is crucial.
In addition, as stated by Eriksen (2011: 6), this system ‘‘is often
associated with the various high-intensity landscape management
systems used by the Arawak, intertwining landscape, history,
myth, subsistence, and travel routes”. In the case of the Paraná
Upper Delta, we believe that earthworking could have served as
a form of ‘‘topographic writing”. Herein, a correlation between
the centrality scores of the mounds and the effort invested in their
construction was observed. Structures located at key intersections
in the transport network could have played the role of topographic
references in the cultural landscape. In this sense, they were prob-
ably associated to the spatial and historical knowledge transmitted
over generations. Even today, both archaeological and modern
mounds function as important reference points for circulating
within the hydrographic transport network. Thus, earthworks that
stand out in terms of size, centrality, settlement stability and asso-
ciation to ancestors could have acted as landmarks and topograms



66 E. Apolinaire, L. Bastourre / Journal of Anthropological Archaeology 44 (2016) 56–68
imbued with historical memory. Moreover, this topograms could
have been related to each other to create topographs, that is, wider
systems of meanings and intertwining stories that are evoked and
recreated by means of observation and movement over the land-
scape (Santos-Granero, 1998). In such way, the space becomes a
storied landscape through the active social interaction of people
with their environment over time, and this interaction becomes
an integral part of social knowledge and identity (Langley, 2013).

When considering the sites network (instead of the hydro-
graphic system as a whole), we observe a positive correlation
between mound size and both centrality measures. That is, the lar-
gest mounds are located at places that not only provide a better
access to other sites but also favour the control over communica-
tions between other settlements. As mentioned above, mound con-
struction required the mobilization of resources and organization
of communal labour. Such organization, in turn, involves the exis-
tence of people with the capacity to interpellate other members of
the society. The extent to which social actors can occupy mediation
positions depends, partly, on their location within a communica-
tion network. Then, social hierarchy and power depend on the
way in which social actors are positioned and connected to each
other (Mizoguchi, 2009). It has been widely recognized that acquir-
ing, exercising and challenging power are relational processes in
which goods, knowledge, symbols and other resources essential
for enacting power in all forms, flow differentially. Settlement
positions within interaction networks influence the kinds of roles
its residents can play in political structures at regional scales
(Schortman, 2014). We believe that, in our study case, these topo-
logical positions could be a key factor for the existence of ‘‘leaders”
who, as attested by archaeological and etnohistorical evidence,
were characterized by the possession of prestige goods and the
ability to create alliances and mobilize resources for exchange,
construction, war, etc. Most probably, these individuals were
located in strategic positions that allowed them to, on the one
hand, better access the hydrographic network and its resources
(HTN) and, on the other, enhanced their control over the flow of
goods, information and people within the sites network (ASN).
All things considered, central network positions would have facil-
itated the mobilization of resources (including people and earth
for erecting mounds) and the control of foreign materials circula-
tion. In effect, the fact that the most central locations within the
sites network are those presenting the highest concentration of
such materials has been noted.
9. Conclusions and future directions

The methodology resorted to for this article combines GIS tools
and concepts derived from graph theory to analyze geographical
and regional archaeological data. Through this approach, we assess
the topological structure of the Paraná Upper Delta fluvial system
and provide an insight as to how the regional archaeological data
behave in relation to network models derived from geographical
information. Unlike other archaeological approaches to transport
systems, where the considered networks are anthropic (e.g. the
Inca road, Antonine Itinerary), the geographical configuration of
the study area allows us to consider the natural hydrographic sys-
tem as a transport network and analyze the position of archaeolog-
ical sites within it. Thus, in the HTN, the transportation network
operates as an independent variable with respect to the settlement
system.

From these studies, the following observations were made: (1)
the sites are not randomly distributed but mainly located in areas
of high closeness (i.e. accessibility); (2) there is a positive correla-
tion between mound size and centrality values (both in the hydro-
graphic network and in the sites network); and (3) foreign
materials are usually found in sites of high centrality (in relation
to the sites network). Based on these results, we propose that (1)
the topology of the hydrographic network conditioned the location
of settlements, (2) the most prominent places are central in terms
of the cultural landscape and (3) the position within interaction
networks conditioned the existence of hierarchical relationships
and the chances to accumulate foreign goods.

While this work represents a first insight into the nature of the
settlement system in the Paraná Upper Delta as regards a network
approach, we believe that the research herein conducted can be
further extended in several ways; firstly, through the application
of other centrality measures as well as the analysis of further net-
work properties and, secondly, by covering a larger spatial scale
and analyzing, by way of more detailed chronologies, variations
in network configuration along successive time slices. It would also
be interesting to compare the hydrographic network model pre-
sented in this paper with a network model drawn from satellite
images corresponding to Paraná River exceptional floods. For those
circumstances, connectivity is expected to have substantially
broadened, thus providing a new scenario to be explored. Finally,
we believe that the models proposed herein allow us to make pre-
dictions regarding site locations that may be tested in future
archaeological surveys.
Acknowledgements

This study was funded by the Consejo Nacional de Investiga-
ciones Científicas y Técnicas of Argentina (CONICET), by way of
Ph.D scholarships and research grants (PIP-CONICET 1282), and
the Agencia Nacional de Promoción Científica y Tecnológica
(ANPCyT) (PICT-0665). We are grateful to Gustavo Politis, Mariano
Bonomo and Carola Castiñeira Latorre for the critical reading of this
paper. Authors are responsible for the views expressed in the
paper.
References

Acuto, F., 2013. ¿Demasiados Paisajes?: Múltiples teorías o múltiples subjetividades
en la arqueología del paisaje. Anuario de Arqueología 5, 31–50.

Apolinaire, E., Castiñeira Latorre, C., Bonomo, M., Politis, G., 2015. Estrategias
prehispánicas de ocupación del espacio en áreas de vulnerabilidad
hidrometeorológica: el caso del Delta del Paraná. Actas del 6�Congreso de
Cuaternario y Geomorfología. CADIC-CONICET, Ushuaia.

Apolinaire, E., Bastourre, L., Costa Angrizani, R., 2016. Arqueología de las tierras altas
de Entre Ríos: primeros resultados de las prospecciones en el interior del
departamento Gualeguay. Intersecciones en Antropología 17 (1), 91–107.

Bafna, S., 2003. Space Syntax. A brief introduction to its logic and analytical
techniques. Environ. Behav. 35 (1), 17–29. http://dx.doi.org/10.1177/
0013916502238863.

Balée, W., Erikson, C., 2006. Time and Complexity in Historical Ecology. Studies in
Neotropical Lowlands. Columbia University Press, New York. http://dx.doi.org/
10.1525/aa.2007.109.2.366.

Barabási, A.-L., Albert, R., 1999. Emergence of scaling in random networks. Science
286, 509–512.

Bastourre, L., 2015. Estudios arqueofaunísticos en el Delta Superior del Paraná: el
sitio los Tres Cerros 1. Revista Chilena de Antropología 30 (2), 109–115.

Bender, B., 1993. Introduction: landscape – meaning and action. In: Bender, B. (Ed.),
Landscape: Politics and Perspectives. Berg, Oxford, pp. 49–84.

Bermejo Tirado, J., 2009. Leyendo los espacios: una aproximación crítica a la sintaxis
espacial como herramienta de análisis arqueológico. Arqueología de la
arquitectura 6, 47–62. http://dx.doi.org/10.3989/arqarqt.2009.09004.

Bonomo, M., 2012. Historia Prehispánica de Entre Ríos. Fundación de Historia
Natural ‘‘Félix de Azara”, Buenos Aires.

Bonomo, M., Politis, G., Gianotti, C., 2011. Montículos, jerarquía social y horticultura
en las sociedades indígenas del Delta del río Paraná (Argentina). Latin Am.
Antiq. 22, 297–333. http://dx.doi.org/10.7183/1045-6635.22.3.297.

Bonomo, M., Politis, G., Castro, J.C., 2010. Primeros resultados de las investigaciones
arqueológicas en el Delta Superior del Paraná y su contribución al atlas
arqueológico de la provincia de Entre Ríos. Folia Histórica del Nordeste 18, 33–
58.

Bonomo, M., Politis, G., Castro, J.C., 2014. Los indígenas de Entre Ríos. In: Manual de
Entre Ríos I. Ministerio de Cultura y Comunicación de la provincia de Entre Ríos.
Editorial de Entre Ríos, Paraná, pp. 45–75.

http://refhub.elsevier.com/S0278-4165(16)30099-X/h0005
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0005
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0005
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0015
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0015
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0015
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0015
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0010
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0010
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0010
http://dx.doi.org/10.1177/0013916502238863
http://dx.doi.org/10.1177/0013916502238863
http://dx.doi.org/10.1525/aa.2007.109.2.366
http://dx.doi.org/10.1525/aa.2007.109.2.366
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0030
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0030
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0035
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0035
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0040
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0040
http://dx.doi.org/10.3989/arqarqt.2009.09004
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0050
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0050
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0050
http://dx.doi.org/10.7183/1045-6635.22.3.297
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0060
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0060
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0060
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0060
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0065
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0065
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0065


E. Apolinaire, L. Bastourre / Journal of Anthropological Archaeology 44 (2016) 56–68 67
Bonomo, M., Blasi, A., 2010. Base regional de recursos líticos del Delta del Paraná.
Estudio petrográfico de artefactos y afloramientos en el sur de Entre Ríos.
Cazadores Recolectores del Cono Sur 4, 17–41.

Brughmans, T., 2013. Thinking through networks: a review of formal network
methods in archaeology. J. Archaeol. Method Theory 20, 623–662. http://dx.doi.
org/10.1007/s10816-012-9133-8.

Brughmans, T., 2014. The roots and shoots of archaeological network analysis: a
citation analysis and rewiew of the archaeological use of formal network
methods. Archaeol. Rev. Cambr. 29 (1), 18–41.

Brughmans, T., Poblome, J., 2016. Pots in space: understanding Roman pottery
distribution from confronting exploratory and geographical network analyses.
In: Barker, E., Bouzarovski, S., Pelling, C., Isaksen, L. (Eds.), New Worlds Out of
Old Texts: Developing Techniques for the Spatial Analysis of Ancient Narratives.
Oxford University Press, Oxford, pp. 255–280. http://dx.doi.org/10.1093/acprof:
oso/9780199664139.003.0011.

Brughmans, T., Keay, S., Earl, G., 2015. Understanding inter-settlement visibility in
Iron Age and Roman Southern Spain with exponential random graph models for
visibility networks. J. Archaeol. Method Theory 22 (1), 58–143. http://dx.doi.
org/10.1007/s10816-014-9231-x.

Cardozo, O., Gómez, E., Parras, M., 2009. Teoría de grafos y Sistemas de Información
Geográfica aplicados al Transporte Público de Pasajeros en Resistencia
(Argentina). Revista Transporte y Territorio 1, 89–111.

Castiñeira, C., Blasi, A., Politis, G., Bonomo, M., del Puerto, L., Huarte, R., Carbonari, J.,
Mari, F., García-Rodríguez, F., 2012. The origin and construction of pre-Hispanic
mounds in the Upper Delta of the Paraná River (Argentina). Archaeol.
Anthropol. Sci. 5, 37–57. http://dx.doi.org/10.1007/s12520-012-0107-2.

Castiñeira, C., Blasi, A., Bonomo, M., Politis, G., Apolinaire, E., 2014. Modificación
antrópica del paisaje durante el Holoceno tardío: las construcciones
monticulares en el Delta Superior del río Paraná. Revista de la Asociación
Geológica Argentina 71 (1), 33–47.

Cavallotto, J., Violante, R., Colombo, F., 2005. Evolución y cambios ambientales de la
llanura costera de la cabecera del río de la Plata. Revista de la Asociación
Geológica Argentina 60, 353–367.

Ceruti, C., 2003. Entidades culturales presentes en la cuenca del Paraná Medio
(margen enterriana). Mundo de Antes 3, 111–134.

Chapman, R., 2003. Archaeologies of Complexity. Routledge, London. http://dx.doi.
org/10.4324/9780203451779.

Collar, A.C.F., 2007. Network theory and religious innovation. Mediterran. Histor.
Rev. 22 (1), 149–162. http://dx.doi.org/10.1080/09518960701539372.

Collar, A., Coward, F., Brughmans, T., Mills, B., 2015. Networks in archaeology:
phenomena, abstraction, representation. J. Archaeol. Method Theory 22 (1), 1–
32. http://dx.doi.org/10.1007/s10816-014-9235-6.

Conkey, M., 1984. To find ourselves: art and social geography of prehistoric hunter
gatherers. In: Schrire, C. (Ed.), Past and Present in Hunter Gatherer Studies. Left
Coast Press, Walnut Creek, pp. 253–276.

Cosgrove, D., 1997. Prospect, perspective and the evolution of the landscape idea.
In: Barnes, T., Gregory, D. (Eds.), Reading Human Geography. Arnould Editors,
London, pp. 324–342.

Coward, F., 2013. Grounding the net: social networks, material culture and
geography in the Epipalaeolithic and Early Neolithic of the Near East
(�21,000–6,000 cal BCE). In: Knappett, C. (Ed.), Network Analysis in
Archaeology: New Approaches to Regional Interaction. Oxford University
Press, Oxford, pp. 247–280. http://dx.doi.org/10.1093/acprof:oso/
9780199697090.003.0011.

Dawson, P., 2002. Space syntax analysis of central inuit snow houses. J.
Anthropol. Archaeol. 21, 464–470. http://dx.doi.org/10.1016/s0278-4165(02)
00009-0.

Dobres, M.A., Robb, J., 2000. Agency in archaeology: paradigm or platitude? In:
Dobres, M.A., Robb, J. (Eds.), Agency in Archaeology. Routledge, London,
pp. 3–17.

Eriksen, L., 2011. Nature and Culture in Prehistoric Amazonia. Using G.I.S. to
reconstruct ancient ethnogenetic processes from archaeology, linguistics,
geography, and ethnohistory. Lund Studies in Human Ecology, vol. 12. Lund
University, Lund.

Fernández de Oviedo y Valdez, G. [1547], 1851. Historia general y natural de las
Indias, 2 (1). Imprenta de la Real Academia de la Historia, Madrid.

Freeman, L.C., 1977. A set of measures of centrality based on betweenness.
Sociometry 40 (1), 35–41. http://dx.doi.org/10.2307/3033543.

Friedkin, N., 1991. Theoretical foundations for centrality measures. Am. J. Sociol. 96
(6), 1478–1504. http://dx.doi.org/10.1086/229694.

Gamble, C., 1998. Palaeolithic society and the release from proximity: a network
approach to intimate relations. World Archaeol. 29 (3), 426–449. http://dx.doi.
org/10.1080/00438243.1998.9980389.

Gianotti, C., Bonomo, M., 2013. De montículos a paisajes: procesos de
transformación y construcción de paisajes en el sur de la cuenca del Plata.
Comechingonia 17, 129–163.

Giddens, A., 1984. The Constitution of Society: Outline of the Theory of
Structuration. University of California Press, Berkeley.

Graham, S., 2006. Networks, agent-based models and the Antonine itineraries:
implications for Roman archaeology. J. Mediterran. Archaeol. 19 (1), 45–64.
http://dx.doi.org/10.1558/jmea.2006.19.1.45.

Golitko, M., Feinman, G., 2015. Procurement and distribution of Pre-Hispanic
Mesoamerican Obsidian 900 BC–AD 1520: a social network analysis. J. Archaeol.
Method Theory 22 (1), 206–247. http://dx.doi.org/10.1007/s10816-014-9211-1.

Hage, P., Harary, F., 1983. Structural Models in Anthropology. Cambridge University
Press, Cambridge. http://dx.doi.org/10.1017/cbo9780511659843.
Heckenberger, M., Neves, E., 2009. Amazonian archaeology. Ann. Rev. Anthropol. 38,
251–266. http://dx.doi.org/10.1146/annurev-anthro-091908-164310.

Hillier, B., 1996. Space is the Machine: A Configurational Theory of Architecture.
University Press, Cambridge.

Hillier, B., Hanson, J., 1984. The Social Logic of Space. Press Syndicate of the
University of Cambridge, Cambridge.

Hill, J., Santos-Granero, F. (Eds.), 2002. Comparative Arawakan Histories: Rethinking
Language Family and Culture Area in Amazonia. University of Illinois Press,
Urbana-Chicago.

Hornborg, A., 2005. Ethnogenesis, regional integration, and ecology in prehistoric
Amazonia: toward a system perspective. Curr. Anthropol. 46, 589–620. http://
dx.doi.org/10.1086/431530.

Hornborg, A., Hill, J., 2011. Introduction: ethnicity in ancient Amazonia. In:
Hornborg, A., Hill, J. (Eds.), Ethnicity in Ancient Amazonia: Reconstructing
Past Identities from Archaeology, Linguistics, and Ethnohistory. University Press
of Colorado, Boulder, pp. 1–27.

Howey, M., 2011. Multiple pathways across past landscapes: circuit theory as a
complementary geospatial method to least cost path for modeling past
movement. J. Archaeol. Sci. 38 (10), 2523–2535. http://dx.doi.org/10.1016/
j.jas.2011.03.024.

Ingold, T., 1992. Culture and the perception of the environment. In: Croll, E., Parkin,
D. (Eds.), Culture and the Perception of the Environment. Routledge, London, pp.
39–56.

Ingold, T., 1997. The picture is not the terrain. Maps, paintings and the dwelt-in
world. Archaeol. Dialog. 4 (1), 29–31. http://dx.doi.org/10.1017/
s1380203800000866.

Iriondo, M., 2004. The littoral complex at the Paraná mouth. Quatern. Int. 114, 143–
154. http://dx.doi.org/10.1016/s1040-6182(03)00049-1.

Iriondo, M., Kröhling, D., 2008. Cambios ambientales en la cuenca del Uruguay
(desde el Presente hasta dos millones de años atrás). Universidad Nacional del
Litoral, Santa Fe.

Irwin-Williams, C., 1977. A network model for the analysis of prehistoric trade. In:
Earle, T.K., Ericson, J. (Eds.), Exchange Systems in Prehistory. Academic Press,
New York, pp. 141–151.

Isaksen, L., 2007. Network analysis of transport vectors in Roman Baetica. In: Clark,
J.T., Hagenmeister, E.M. (Eds.), Digital Discovery: Exploring New Frontiers in
Human Heritage, Proceedings of the 34th CAA Conference, Fargo, 2006.
Archaeolingua, Budapest, pp. 76–87.

Isaksen, L., 2008. The application of network analysis to ancient transport
geography: a case study of Roman Baetica. Dig. Medie. 4, URL <http://www.
digitalmedievalist.org/journal/4/isaksen> (last access: October 28, 2015).

Jenkins, D., 2001. A network analysis of Inka roads, administrative centers, and
storage facilities. Ethnohistory 48 (4), 655–687. http://dx.doi.org/10.1215/
00141801-48-4-655.

Knappett, C., 2013. Introduction: why networks? In: Knappett, C. (Ed.), Network
Analysis in Archaeology. New Approaches to Regional Interaction. Oxford
University Press, Oxford, pp. 3–16. http://dx.doi.org/10.1017/
s0003598x0009774x.

Knappett, C., Evans, T., Rivers, R., 2008. Modelling maritime interaction in the
Aegean Bronze Age. Antiquity 82 (318), 1009–1024.

Langley, M., 2013. Storied landscapes makes us (Modern) Human: landscape
socialisation in the Palaeolithic and consequences for the archaeological record.
J. Anthropol. Archaeol. 32, 614–629. http://dx.doi.org/10.1016/
j.jaa.2013.10.001.

Lopes de Sousa, P. [1530-1532], 1839. Diario da Navegação de Pero Lopes de Sousa
(de 1530 a 1532). Typographía da Sociedade Propagadora dos Conhecimentos,
Lisboa.

McRae, B., 2006. Isolation by resistance. Evolution 60 (8), 1551–1561. http://dx.doi.
org/10.1554/05-321.1.

McRae, B., Beier, P., 2007. Circuit theory predicts gene flow in plant and animal
populations. Proc. Natl. Acad. Sci. 104 (50), 19885–19890. http://dx.doi.org/
10.1073/pnas.0706568104.

Medina, J., 1908. El veneciano Sebastián Caboto, al servicio de España y
especialmente de su proyectado viaje á las Molucas por el Estrecho de
Magallanes y al reconocimiento de la costa del continente hasta la
gobernación de Pedrarias Dávila, 2. Imprenta y Encuadernación Universitaria,
Santiago de Chile.

Milana, J., Kröhling, D., 2015. Climate changes and solar cycles recorded at the
Holocene Paraná Delta, and their impact on human population. Sci. Rep. 5, 1–8.
http://dx.doi.org/10.1038/srep12851.

Mizoguchi, K., 2009. Nodes and edges: a network approach to hierarchisation and
state formation in Japan. J. Anthropol. Archaeol. 28 (1), 14–26. http://dx.doi.org/
10.1016/j.jaa.2008.12.001.

Munson, J., Macri, M., 2009. Sociopolitical network interactions: a case study of the
Classic Maya. J. Anthropol. Archaeol. 28 (4), 424–438. http://dx.doi.org/10.1016/
j.jaa.2009.08.002.

Neves, E., 2001. Indigenous historical trajectories in the upper Rio Negro Basin. In:
McEwan, C., Barreto, C., Neves, E. (Eds.), Unknown Amazon: Culture in Nature in
Ancient Brazil. British Museum Press, London, pp. 266–286.

Paira, A., Drago, E., 2007. Origin, evolution, and types of floodplain water bodies. In:
Iriondo, M., Paggi, J., Parma, M. (Eds.), The Middle Paraná River: Limnology of a
Subtropical Wetland. Springer, Berlin, pp. 53–81.

Peregrine, P., 1991. A graph-theoretic approach to the evolution of Cahokia. Am.
Antiq. 56 (1), 66–75. http://dx.doi.org/10.2307/280973.

Pitts, F.R., 1965. A graph theoretic approach to historical geography. Professional
Geograph. 17 (5), 15–20. http://dx.doi.org/10.1111/j.0033-0124.1965.015_m.x.

http://refhub.elsevier.com/S0278-4165(16)30099-X/h0070
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0070
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0070
http://dx.doi.org/10.1007/s10816-012-9133-8
http://dx.doi.org/10.1007/s10816-012-9133-8
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0080
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0080
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0080
http://dx.doi.org/10.1093/acprof:oso/9780199664139.003.0011
http://dx.doi.org/10.1093/acprof:oso/9780199664139.003.0011
http://dx.doi.org/10.1007/s10816-014-9231-x
http://dx.doi.org/10.1007/s10816-014-9231-x
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0095
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0095
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0095
http://dx.doi.org/10.1007/s12520-012-0107-2
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0105
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0105
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0105
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0105
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0110
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0110
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0110
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0115
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0115
http://dx.doi.org/10.4324/9780203451779
http://dx.doi.org/10.4324/9780203451779
http://dx.doi.org/10.1080/09518960701539372
http://dx.doi.org/10.1007/s10816-014-9235-6
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0135
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0135
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0135
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0140
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0140
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0140
http://dx.doi.org/10.1093/acprof:oso/9780199697090.003.0011
http://dx.doi.org/10.1093/acprof:oso/9780199697090.003.0011
http://dx.doi.org/10.1016/s0278-4165(02)00009-0
http://dx.doi.org/10.1016/s0278-4165(02)00009-0
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0155
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0155
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0155
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0160
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0160
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0160
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0160
http://dx.doi.org/10.2307/3033543
http://dx.doi.org/10.1086/229694
http://dx.doi.org/10.1080/00438243.1998.9980389
http://dx.doi.org/10.1080/00438243.1998.9980389
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0185
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0185
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0185
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0190
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0190
http://dx.doi.org/10.1558/jmea.2006.19.1.45
http://dx.doi.org/10.1007/s10816-014-9211-1
http://dx.doi.org/10.1017/cbo9780511659843
http://dx.doi.org/10.1146/annurev-anthro-091908-164310
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0215
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0215
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0220
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0220
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0225
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0225
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0225
http://dx.doi.org/10.1086/431530
http://dx.doi.org/10.1086/431530
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0235
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0235
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0235
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0235
http://dx.doi.org/10.1016/j.jas.2011.03.024
http://dx.doi.org/10.1016/j.jas.2011.03.024
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0245
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0245
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0245
http://dx.doi.org/10.1017/s1380203800000866
http://dx.doi.org/10.1017/s1380203800000866
http://dx.doi.org/10.1016/s1040-6182(03)00049-1
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0260
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0260
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0260
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0265
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0265
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0265
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0270
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0270
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0270
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0270
http://www.digitalmedievalist.org/journal/4/isaksen
http://www.digitalmedievalist.org/journal/4/isaksen
http://dx.doi.org/10.1215/00141801-48-4-655
http://dx.doi.org/10.1215/00141801-48-4-655
http://dx.doi.org/10.1017/s0003598x0009774x
http://dx.doi.org/10.1017/s0003598x0009774x
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0290
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0290
http://dx.doi.org/10.1016/j.jaa.2013.10.001
http://dx.doi.org/10.1016/j.jaa.2013.10.001
http://dx.doi.org/10.1554/05-321.1
http://dx.doi.org/10.1554/05-321.1
http://dx.doi.org/10.1073/pnas.0706568104
http://dx.doi.org/10.1073/pnas.0706568104
http://dx.doi.org/10.1038/srep12851
http://dx.doi.org/10.1016/j.jaa.2008.12.001
http://dx.doi.org/10.1016/j.jaa.2008.12.001
http://dx.doi.org/10.1016/j.jaa.2009.08.002
http://dx.doi.org/10.1016/j.jaa.2009.08.002
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0335
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0335
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0335
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0340
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0340
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0340
http://dx.doi.org/10.2307/280973
http://dx.doi.org/10.1111/j.0033-0124.1965.015_m.x


68 E. Apolinaire, L. Bastourre / Journal of Anthropological Archaeology 44 (2016) 56–68
Pitts, F., 1979. The medieval river trade network of Russia revisited. Soc. Networks 1
(3), 285–292. http://dx.doi.org/10.1016/0378-8733(78)90025-4.

Politis, G., Bonomo, M., Castiñeira, C., Blasi, A., 2011. Archaeology of the Upper Delta
of the Paraná River (Argentina): mound construction and anthropic landscapes
in the Los Tres Cerros locality. Quatern. Int. 245, 74–88. http://dx.doi.org/
10.1016/j.quaint.2011.02.007.

Politis, G., Bonomo, M., 2012. La entidad arqueológica Goya-Malabrigo (Ríos Paraná
y Uruguay) y su filiación Arawak. Boletín de la Sociedade de Arqueologia
Brasileira (SAB) 25 (1), 10–46.

Ramírez, L. [1528], 2007. Carta de Luis Ramírez a su padre desde el Brasil (1528):
orígenes de lo ‘real maravilloso’ en el Cono Sur. URL: <http://parnaseo.uv.es/
Lemir/Textos/Ramirez.pdf> (last access: October 28, 2015).

Rapoport, A., 1990. The Meaning of the Built Environment: A Non-Verbal
Communication Approach. University of Arizona, Tucson.

Richards-Rissetto, H., Landau, K., 2014. Movement as a means of social (re)
production: using GIS to measure social integration across urban landscapes. J.
Archaeol. Sci. 41, 365–375. http://dx.doi.org/10.1016/j.jas.2013.08.006.

Rostain, S., 2010. Pre-Columbian earthworks in coastal Amazonia. Diversity 2 (3),
331–352. http://dx.doi.org/10.3390/d2030331.

Rothman, M., 1987. Graph theory and the interpretation of regional survey data.
Paléorient 13 (2), 73–91. http://dx.doi.org/10.3406/paleo.1987.4430.

Sánchez, J., Colobig, M., Zucol, A., Politis, G., Bonomo, M., Castiñeira, C., 2013.
Primeros resultados sobre el uso prehispánico de los vegetales en el sitio
arqueológico Los Tres Cerros 1 (Victoria, Entre Ríos, Argentina): análisis del
registro biosilíceo. Darwiniana 1 (2), 201–219.

Santa Cruz, A., [1540], 1918. Islario general de todas las islas del mundo. Imprenta
del Patronato de Huérfanos de intendencia é intervención Militares, Caracas.

Santos-Granero, F., 1998. Writing history into the landscape: space, myth, and ritual
in contemporary Amazonia. Am. Ethnol. 25, 128–148.

Santos-Granero, F., 2002. The Arawakan Matrix: Ethos, Language, and History in
Native South América. In: Hill, J., Santos-Granero, F. (Eds.), Comparative
Arawakan Histories: Rethinking Language Family and Culture Area in
Amazonia. University of Illinois Press, Urbana-Chicago, pp. 25–50.
Scabuzzo, C., Ramos Van Raap, A., Bonomo, M., Politis, G., 2015. Bioarchaeological
study of the Los Tres Cerros site 1 (Upper Delta of the Paraná River, Entre Ríos,
Argentina). Boletim do Museu Paraense Emílio Goeldi. Ciências Humanas 10 (2),
509–536.

Schmidl, U. [1567], 1980. Derrotero y viaje a España y a las Indias. Espasa-Calpe,
Buenos Aires.

Schortman, E., 2014. Networks of power in archaeology. Ann. Rev. Anthropol. 43,
167–182. http://dx.doi.org/10.1146/annurev-anthro-102313-025901.

Serrano, A., 1950. Los primitivos habitantes de Entre Ríos. Ministerio de Educación,
Paraná.

Sevtsuk, A., Mekonnen, M., 2012. Urban network analysis: a new toolbox for ArcGIS.
Revue Internationale de Géomatique 2, 287–305.

Souza, J., Corteletti, R., Robinson, M., Iriarte, J., 2016. The genesis of monuments:
resisting outsiders in the contested landscapes of southern Brazil. J. Anthropol.
Archaeol. 41, 196–212. http://dx.doi.org/10.1016/j.jaa.2016.01.003.

Terrell, J.E., 2010. Language and material culture on the Sepik Coast of Papua New
Guinea: using social network analysis to simulate, graph, identify, and analyze
social and cultural boundaries between communities. J. Island Coast. Archaeol.
5 (1), 3–32. http://dx.doi.org/10.1080/15564890903142891.

Thayn, J., Sampeck, K., Spaccapaniccia, M., 2016. Refining Hernando de Soto’s Route
Using Electric Circuit Theory and CircuitScape. Profession. Geograph. 1–8.
http://dx.doi.org/10.1080/00330124.2015.1124787.

Thomas, J., 2001. Archaeologies of place and landscapes. In: Hodder, I. (Ed.),
Archaeological Theory Today. Polity Press, Cambridge, pp. 165–186.

Tilley, C., 1993. Art, architecture, landscape [Neolithic Sweden]. In: Bender, B. (Ed.),
Landscape: Politics and Perspectives. Berg, Oxford, pp. 49–84.

Wallis, W., 2007. A Beginner’s Guide to Graph Theory. Springer Science & Business
Media, Boston.

Wasserman, S., Faust, K., 1994. Social Network Analysis: Methods and Applications.
Cambridge University Press, Cambridge.

Wilson, R., 2014. History of graph theory. In: Gross, J., Yellen, J., Zhang, P. (Eds.),
Handbook of Graph Theory, second ed. Taylor & Francis, New York,
pp. 31–51.

http://dx.doi.org/10.1016/0378-8733(78)90025-4
http://dx.doi.org/10.1016/j.quaint.2011.02.007
http://dx.doi.org/10.1016/j.quaint.2011.02.007
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0365
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0365
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0365
http://parnaseo.uv.es/Lemir/Textos/Ramirez.pdf
http://parnaseo.uv.es/Lemir/Textos/Ramirez.pdf
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0375
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0375
http://dx.doi.org/10.1016/j.jas.2013.08.006
http://dx.doi.org/10.3390/d2030331
http://dx.doi.org/10.3406/paleo.1987.4430
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0395
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0395
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0395
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0395
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0405
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0405
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0410
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0410
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0410
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0410
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0415
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0415
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0415
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0415
http://dx.doi.org/10.1146/annurev-anthro-102313-025901
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0430
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0430
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0435
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0435
http://dx.doi.org/10.1016/j.jaa.2016.01.003
http://dx.doi.org/10.1080/15564890903142891
http://dx.doi.org/10.1080/00330124.2015.1124787
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0455
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0455
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0460
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0460
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0465
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0465
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0470
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0470
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0475
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0475
http://refhub.elsevier.com/S0278-4165(16)30099-X/h0475

	Nets and canoes: A network approach to the pre-Hispanic settlement system in the Upper Delta of the Paraná River (Argentina)
	1 Introduction
	2 Graph theory and network analysis in archaeology
	3 Waterways, graphs and centrality in the Paraná Delta
	4 Environmental framework
	5 Archaeological framework
	6 Material and methods
	7 Results
	8 Discussion
	9 Conclusions and future directions
	Acknowledgements
	References