transportation corridors and political evolution in...

Available online at www.sciencedirect.com

Journal of Anthropological Archaeology 26 (2007) 607–629

www.elsevier.com/locate/jaa

Transportation corridors and political evolution in highlandMesoamerica: Settlement analyses incorporating GIS

for northern Tlaxcala, Mexico

David M. Carballo a,*, Thomas Pluckhahn b,1

a Cotsen Institute of Archaeology at UCLA, A210 Fowler Building, Box 951510, Los Angeles, CA 90095-1510, USAb Department of Anthropology, University of South Florida, 4202 E. Fowler Ave., SOC 107, Tampa, FL 33620-8100, USA

Received 10 October 2006; revision received 2 May 2007Available online 12 June 2007

Abstract

Investigations of the functional utility and social elaboration of natural transportation corridors contribute to generatingmore comprehensive understandings of complex macroregional phenomena such as political evolution. In this article, high-land central Mexican transportation corridors are analyzed through GIS applications and a reconsideration of settlement datafrom an important corridor in northern Tlaxcala. The location of the study region, within the corridor connected to the north-eastern Basin of Mexico, makes it particularly relevant for examining interregional exchanges and the transformation of amore rural landscape adjacent to core areas of urbanization and state political expansion. Conclusions drawn from the studyare largely complementary to the interpretations made by the surveyors [Merino Carrion, B.L., 1989. La Cultura Tlaxco. SerieArqueologica, Coleccion Cientıfica 174, Instituto Nacional de Antropologıa e Historia, Mexico City], but provide an updatedassessment of prehispanic political evolution ca. 900 BC–AD 1519. Specifically, the study tracks the evolution of early cere-monial centers, Teotihuacan’s territorial expansion through the region, political balkanization following the collapse of Teo-tihuacan, and the establishment of small, independent polities encountered by the Spanish. New contributions includequantitative measures relevant to assessing the systemic integration of the region and the social value of the corridor duringdifferent chronological phases, elucidating the differing expansionary trajectories of Teotihuacan and the Aztec Empire.� 2007 Elsevier Inc. All rights reserved.

Keywords: State expansion; Settlement analysis; GIS; Cost path; Transportation corridor; Mesoamerica

Highland peoples across the world negotiate therugged terrain around them through natural trans-portation corridors that facilitate the movement ofindividuals, goods, and ideas within highland envi-ronments and between highlands and lowlands.

0278-4165/$ - see front matter � 2007 Elsevier Inc. All rights reserveddoi:10.1016/j.jaa.2007.05.001

* Corresponding author. Fax: +1 310 206 4723.E-mail addresses: [email protected] (D.M. Carballo), tpluckha

@cas.usf.edu (T. Pluckhahn).1 Fax: +1 813 974 2668.

Such corridors expedite the cumulative processesof cultural exchange, becoming important nodes inpolitical evolution and frequent targets for controlby dominant polities. The highland civilizations ofprehispanic Mesoamerica developed particularstrategies for establishing interregional contactsthrough inter-montane corridors, without the assis-tance of pack animals or wheeled vehicles. Theseexchanges included economic and ideologicalrelations between largely autonomous Formative

.

mailto:[email protected]

mailto:tpluckha @cas.usf.edu

mailto:tpluckha @cas.usf.edu

Fig. 1. Chronology for northern Tlaxcala, with PANT phasesbased on Merino Carrion (1989).

608 D.M. Carballo, T. Pluckhahn / Journal of Anthropological Archaeology 26 (2007) 607–629

polities (ca. 1200 BC–AD 100), and the more exten-sive political entanglements centered on the urbanstates and empires of the Classic and Postclassicperiods (ca. AD 100–1519) (Fig. 1).

Fig. 2. Central Mexico with stud

In this study, we incorporate settlement analysesassisted by GIS applications and new data from recentexcavations at the largest Formative period center innorthern Tlaxcala, a region which straddles an impor-tant transportation corridor leading from the Basin ofMexico to points east and south (Fig. 2). We considersocial processes of varying scales, including internalpolitical dynamics and the effects of multiregionalphenomena such as urbanization and state expansioncentered in regions adjacent to the study area. Draw-ing on multiple lines of evidence, our study illuminatespolitical processes that affected prehispanic centralMexico at several levels, with general implicationsfor the study of ancient cultural geography, polity for-mation, and political expansion.

Transportation corridors and political evolution

Archaeologists have identified several importantcommonalities among cases of political evolution inpast societies. Significant among these is the relation-ship between decision-making infrastructure, trans-portation capabilities, and polity size. Whereas

y region and selected sites.

D.M. Carballo, T. Pluckhahn / Journal of Anthropological Archaeology 26 (2007) 607–629 609

polities lacking internally specialized bureaucraticinstitutions that could effectively administer distantcommunities were restricted to control radii consis-tent with a one day journey from the political capital,those that evolved or adopted such institutions werecapable of consolidating much larger territories intoexpansive states or empires (Spencer, 1990, 1998;Wright, 1984, 1986). Several historically and archae-ologically well-documented cases support the propo-sition that territorial expansion provided the initialstimuli for the evolution of bureaucratic control insti-tutions (Algaze, 1993; Flannery, 1999; Tilly, 1990;Wright, 2006). In these cases state political institu-tions followed cycles of sustained conflict as a meansof initiating tributary payments and/or securingaccess to distant, valued resources.

Other dynamics in the processes of political evo-lution are more temporally and spatially contingent.They include the relative degrees of administrativecomplexity of neighboring communities and socie-ties; the myriad goals and motivations of individualsalong the sociopolitical and settlement spectrum;regional distributions of resources, and naturaland technological factors involved in accessingthem. The effective balancing of such considerationsfor particular archaeological sequences requiresmulti-scalar approaches that successfully accountfor observable patterning at the community, regio-nal, and macroregional levels.

Comparative multi-scalar approaches are oftenuseful in illustrating general trends that are relevantto investigating political evolution and to anthropo-logical model building. For instance, in discussingearly Andean state expansion Schreiber (1987) andStanish (2003) note the importance of the particularhistorical conditions separating their cases from laterInca imperial expansion. At a broad level, however,both authors accentuate the significance of lower pre-existing complexity to the expansionary strategiespursued by Wari and Tiwanaku leaders, respectively,and the manner in which these differed from Incaimperial policies (see also Jennings and Craig, 2001).

Interest in the dialectic between political powerand resistance has moved to the forefront in archae-ology as a result of the increased attention focusedon past human agency, and the proliferation of exca-vations away from ancient political capitals. We echothe sentiments of a number of researchers regardingthe necessity of considering the array of motivationsand/or constraints on the inhabitants of expansion-ary polities, as well as those of the smaller polities thatwere subject to, or threatened by such expansion

(e.g., Falconer and Savage, 1995; Gledhill, 1988;Stein, 2002). Resistance strategies and the differingmotivations of political subordinates are often elu-sive in cases lacking textual evidence; however,robust archaeological datasets provide multipledimensions for considering their roles more carefully.For example, archaeologists working in Oaxaca,Mexico, combine multiple lines of evidence in docu-menting the more erratic, less incremental expansionof the Monte Alban state, the prioritization by stateleaders of controlling key transportation corridors,and the varied reactions of neighboring communitiesorganized at differing levels of complexity (Balkan-sky, 2002; Marcus and Flannery, 1996, pp. 195–207; Spencer and Redmond, 2006).

Territorial expansion and strategies of controlpursued by political elites were frequently shapedby the rapidity by which goods, armies, and informa-tion could move across the surrounding landscape.What constitutes a transportation corridor may varywidely among societies (i.e., rivers, straights, moun-tain passes). Their desirability is related to their prox-imity to valued resources and exchange networks,existing transportation technologies, and other con-tingent factors such as those mentioned above.Because of their importance for channeling humanexchanges, the identification of desired transporta-tion corridors, and the explanation of changes in theirusage over time, significantly contributes to under-standing macroregional phenomena such as politicalevolution (Algaze, 1993; Burghardt, 1971; Hirth,1978a; Sherratt, 2000). Contemporary GIS applica-tions assist in the accurate recognition of importantcorridors, and allow us to reanalyze settlement pat-terning within and surrounding them from preexist-ing survey projects, providing another compellingline of evidence for elucidating larger-scale processes.

Political evolution and the Mexican altiplano

Although likely apocryphal, the often repeatedstory of Cortes crumpling a piece of paper before Car-los I as a visual representation of New Spain nicelyillustrates that its contoured landscape has long beenrecognized as circumscribing human movements(Melville, 2000, pp. 215–216; see also Cortes, 1519/1986a, p. 29). The wide basins, valleys, and plains ofthe Mesoamerican highlands are separated by volca-nic ranges often rising 500–2000 m from their floorsand channeling movement through the corridorsbetween them. Coupled with the exclusive relianceon human portage for overland movements of goods,

2 Garcıa Cook (1972) described spacing between surveyors as nothaving exceeded 200 m, yet does not mention the most commonlyused interval. When sites were encountered, their locations wererecorded on 1:50,000 scale aerial photos. Intensive site walkingestablished their boundaries, chronological classifications werebased on surface artifacts gathered by spot collections, andenvironmental, architectural, and artifactual information wasrecorded for each site (forms are reproduced in Garcıa Cook, 1972and Merino Carrion, 1989). The PANT surveyors incorporated sitesidentified by Snow, who had published the results of his partiallyoverlapping survey prior the initiation of the PANT project (Snow,1966, 1969). They added to his list and recorded over twice thedensity of sites in areas where the two surveys overlap (approxi-mately 0.12/km2 and 0.05/km2, respectively). Total site density isover nine-times lower for the PANT region when compared to thedensity recorded for 3500 km2 in the Basin of Mexico (approxi-mately 1.10/km2; Sanders et al., 1979, p. 185). However, we contendthat this discrepancy reflects the different demographic histories ofthese two regions, rather than any serious variability in the siterecovery rates of the two surveys.

The estimated 90% survey coverage and 95% settlement recoverycited by Merino Carrion are not elaborated on further. We assumethat the unsurveyed portions include areas that were covered inmodern habitation during the mid-1970s, and the upper regions ofthe Sierra Tlaxco and Sierra Terrenate, which provide a naturalboundary for the survey region and state of Tlaxcala runningnorthwest to southeast. This boundary is meaningful in terms ofcontaining settlement, and the study region as a whole is generally acoherent unit consisting of the Continental Divide, with the upperportions of the Zahuapan and Apulco drainages. Nevertheless, thewestern and southern boundaries of the survey region are arbitrarilydefined, and simply represent the final swath of the greater Tlaxcala–Puebla region to be surveyed by Garcıa Cook’s team. Unfortu-nately, data from adjacent areas acquired by Garcıa Cook andINAH Tlaxcala are not available in a format that would allow us toextend our analyses in these directions—the primary data of theformer not being published in an accessible format such as was doneby Merion Carrion, and the latter being chronologically coded intooverly broad Formative, Classic, and Postclassic phases. However,as we argue in our interpretations of the data, we believe that thelarge majority of PANT settlements represent a meaningful unit ofanalysis, except for occasional clusters of sites in the southwesternportion of the study region, which we comment on in the main text.


this ecological setting provided particular structuralconditions within which highland peoples pursuedstrategic economic and political interactions withtheir neighbors and more distant societies. Amongthese, many researchers have emphasized the role oftransportation corridors in interregional exchangeand the expansionary strategies pursued by the lead-ers of highland state polities, including Teotihuacanand the Aztec (Triple Alliance) Empire (e.g., Charl-ton, 1978, 1991; Drennan, 1984; Garcıa Cook andMerino Carrion, 1977, 1996; Hassig, 1985; Hirth,1978a; Litvak King, 1978; Santley and Pool, 1993).These highland polities politically integrated largeportions of central Mexico and leveraged access todesired resource zones elsewhere in Mesoamerica ina variety of ways, all ultimately backed by large,well-organized armies. The multiregional politicalstrategies of the ethnohistorically documented Aztecare much better known to us than is the case for Teo-tihuacan, whose political influence outside the Basinof Mexico has been reconstructed primarily throughsettlement analyses and the presence of Teotihuacanoproducts or styles abroad (compare Millon, 1988;Smith and Hodge, 1994).

Whether considering large expansionistic states orsmaller polities that nevertheless interacted fre-quently with other regions, Mesoamerican scholarsrecognize the importance of multiregional social phe-nomena in stimulating settlement changes in one par-ticular region or sub-region (Balkansky, 2006). Wefollow in this tradition by reconsidering settlementpatterns from northern Tlaxcala within the contextof large-scale developments that likely affectedthem—as was also done by the original surveyorsof the region (Garcıa Cook, 1981; Garcıa Cook andMerino Carrion, 1996; Merino Carrion, 1989). Addi-tionally, we contribute new insights based on excava-tion data and analytical methods not available to thesurveyors at the time their work was undertaken.

Important multiregional phenomena likely to haveaffected communities in northern Tlaxcala include: (1)highland–lowland exchanges in goods and ideas run-ning east–west through the region; (2) the urbaniza-tion of central Mexico during the first few centuriesBC; (3) Teotihuacan’s political/territorial expansionand provincial administration during the first sixcenturies AD; (4) macroregional balkanization intocompeting city-states following the collapse of Teo-tihuacan; and (5) regional resistance to Aztec imperialexpansion prior to the Spanish Conquest. In recon-structing the political evolution and interregionalinteractions of northern Tlaxcalan communities we

gain an improved appreciation for the macroregionalscale of these developments, away from their morearchaeologically conspicuous centers.

The study region and database

The database used in this study was compiledthrough intensive survey in 1972–1975 by the Proyec-to Arqueologico del Norte de Tlaxcala (PANT)under the direction of Angel Garcıa Cook (publishedby Merino Carrion in 1989). The study area covers1200 km2 of northern Tlaxcala and Merino Carrion(1989, pp. 27–30) estimated that 90% of the studyregion was covered by the survey, documentingapproximately 95% of archaeological sites.2


Although the region is slightly more marginal foragriculture than the southern central highlands—with a valley floor over 2500 m.a.s.l. receivingapproximately 80 cm of precipitation a year—its pre-hispanic inhabitants were continuously connected tothe macroregional interaction spheres of greater cen-tral Mexico (Garcıa Cook, 1981). Presently, the rail-road and freeway connecting Mexico City to the portof Veracruz run through the southern portion of thestudy region, offering a modern analogy to its impor-tance as a trade corridor in prehispanic times.

One of the many interesting patterns documentedby Angel Garcıa Cook in his archaeological surveyof the Puebla–Tlaxcala region was the dramaticshift in settlement coinciding with Classic periodurbanization at Teotihuacan and Cholula—whicharose at alternate ends of the region to the north-west and south, respectively (Garcıa Cook, 1981;Garcıa Cook and Merino Carrion, 1996; GarcıaCook and del Carmen Trejo, 1977). Garcıa Cook

Fig. 3. Northern Tlaxcala study regio

characterized northern Tlaxcala as possessing a‘‘Teotihuacan Sphere,’’ consisting of northwesternsites exhibiting strong material affinities to the city,which likely formed the eastern portion of the coreof the polity (see also Millon, 1988). He also desig-nated a ‘‘Teotihuacan Corridor’’ running northwestto southeast through northern Tlaxcala, consistingof sites exhibiting weaker material affinities to thecity, but which differed significantly from the localTenanyecac phase culture (AD 100–600) found out-side the natural corridor of transportation (see alsoVega Sosa, 1981).

The PANT survey region consists of this naturaltransportation corridor—which we refer to hereaf-ter as the Tlaxcala Corridor, as we discuss it withinthe context of developments preceding and follow-ing Teotihuacan—and the mountain ranges thatenclose it to the north (Fig. 3). The eastern portionof the study region is located at the opening to a sec-ond natural transportation corridor, extending

n with transportation corridors.


southwest to northeast from the Tlaxcala Corridorinto northern Puebla through these mountainranges, between the western Sierra of Tlaxco andthe eastern Sierra of San Nicolas Terrenate. Thiscanyon was cut by the Apulco River and is bestknown to central Mexican scholars as Cortes’slikely point of entry into independent Tlaxcala,arriving from the town of Ixtacamaxtitlan en routeto Tenochtitlan (Angulo Ramırez, 2004, p. 15;Cortes, 1519/1986b, pp. 56–58; Davies, 1968, p.67; Garcıa Martınez, 2001; Munoz Camargo,1584/2000, pp. 104–105). We will refer to this sec-ond transportation corridor as the Apulco Canyon.

Research methods

Our analyses incorporated a multi-scalarapproach: (1) quantifying of the utility of the Tlax-cala Corridor for interregional transportation rela-tive to other potential corridors; (2) examiningspatial distributions between sites in the PANT sur-vey region during different chronological phases toassess potential hierarchical relationships betweencontemporaneous sites, and site locations relativeto the Tlaxcala Corridor; and (3) calculating rank-size distributions for each phase in the region usingthe A coefficient with bootstrapped error ranges fol-lowing the work of Drennan and Peterson (2004).

We met these objectives through the constructionof a GIS database for central Mexico, and bymanipulating and modifying the PANT data so thatit could be incorporated into the database and rank-size analyses. In constructing the GIS database weused a surface elevation mosaic with finishedthree-arc-second raster data (90 m elevations)acquired from the Shuttle Radar Topography Mis-sion (www.seamless.usgs.gov). This step allowedfor the production of detailed topographic mapson which spatially referenced data could be pro-jected and analyzed.

The least cost path analysis involved calculatingthe least accumulative cost of travel between fea-tures for each cell in the raster dataset, taking intoaccount surface distance, horizontal-cost factors,and vertical-cost factors. Specifically, we utilizedthe hiker function developed by Tobler (1993) inconjunction with the Pathdistance tool in ArcInfo(ESRI, Inc.) to create an anisotropic friction surfacerepresenting the time in hours of foot travel fromvarious features, as well as a second surface repre-senting the most efficient direction of travel. Thesesurfaces were then used with the cost path tool in

ArcGIS (ESRI, Inc.) to plot the routes with the leasttravel time. Our methods are similar to those usedby Jennings and Craig (2001), but we were inter-ested in calculating travel time as a means of defin-ing important transportation corridors, rather thanin defining political boundaries.

It should be noted that this type of least cost anal-ysis has a number of potential errors, owing to lim-itations on the manner in which the accumulatedcosts and directions are calculated (see Conollyand Lane, 2006, pp. 221–224; Wheatley and Gillings,2002, pp. 154–158). Nevertheless, we believe that theleast cost path analysis provides a reasonably accu-rate relative estimate of the difficulty of foot travelacross the landscape, as well as the most likely routesof travel for the region’s inhabitants (see Aldender-fer, 1998; Kantner, 1997). Physiographic factorssuch as rivers and vegetative cover have not beenconsidered in our analysis. However, the narrow riv-ers and streams of the region provide relatively fewimpediments, and vegetation is most dense on theslopes of its volcanic ranges—only augmenting therelative costs of slope travel.

Our features included sites from the PANT sur-vey, other important central Mexican sites outsideof the survey region, Monte Alban, and four pointslocated along the Gulf Coast spaced at 100 kmintervals. We chose arbitrary points along the coastrather than actual archaeological sites because wewere more interested in gauging central Mexicanaccess to coastal resources over the entire sequence,rather than particular site connections during cer-tain phases.

We incorporated the PANT settlement datareported in the maps and tables published by Mer-ino Carrion (1989) into the GIS database as layerscorresponding to each chronological phase. Theomission of six sites registered in the data tablesbut not on the map of sites (Merino Carrion,1989, Fig. 6) resulted in a total of 126 sites for allphases. We do not consider the loss of these fewrural sites to significantly affect our analyses andinterpretations. Our most noteworthy modificationof the data was to increase the size estimates for siteT-491, which affects the rank-size analysis. This site,La Laguna, has recently been investigated by Carb-allo and Aleksander Borejsza, as part of projectsdirected by Richard Lesure (2003–2004) and Carb-allo (2005–2006). It was first documented in thearchaeological literature by Snow (1966, 1969),who registered it as the largest Formative period sitein his 1500 km2 survey of central and northern

http://www.seamless.usgs.gov


Tlaxcala, which partially overlaps the PANT surveyregion. Carballo mapped the site in 2004–2005. Themapping program, excavations, and numerousinspections of surface artifact densities result inour revised size estimates for La Laguna as coveringapproximately 100 ha during the Tezoquipan phaseoccupation. Our revised estimate of 66 ha as thecommunity’s size during the Texoloc phase is con-servative, and represents a middle ground betweenthe extent of the site at its height and where Texolocphase features have been excavated. It is likely thatthe PANT surveyors underestimated the size of LaLaguna (listing it as 30 ha during both phases)because of its positioning within the saddle betweenthree surrounding hills, which has resulted in sub-stantial colluvial blanketing, leaving far fewer arti-facts visible on the surface than is usual forhighland Mexico.

Several analyses were undertaken without furtherreclassification of the data, while certain spatialanalyses and the phase-by-phase settlement mapsemployed a hierarchical ranking scheme based onthe number of ceremonial structures documentedfor each site. We consider the relative frequenciesof ceremonial architecture at contemporaneous siteswithin a survey region to be a reliable index of hier-archical relationships between past communities(see Balkansky, 1998; Blitz, 1999; Steponaitis,1978). Mounds are conspicuous features on the cen-tral Mexican landscape, making them easier to reli-ably quantify than are site boundaries defined bysherd scatter, which are more likely to have beenaffected by post-depositional processes such asplowing and slope erosion.

Our site classifications based on ceremonialarchitecture created the following ordinal catego-ries: (Tier 4) no visible ceremonial mounds; (Tier3) at least one ceremonial mound; (Tier 2) at leasttwice the number of ceremonial mounds of anythird-tier site for the same phase; and (Tier 1) atleast twice the number of ceremonial mounds ofany second-tier site for the same phase. The motiva-tion for developing this scheme was to create simpli-fied categories that would be useful in examininghow different types of sites (ceremonial centers ofdifferent sizes versus non-ceremonial centers) spa-tially articulated with each other on the regionallandscape. If the quantity of public ceremonialarchitecture is a reliable proxy for the number ofadministrative functions of a community, then cere-monial centers should generally be more dispersedover the landscape so that they are more optimally

located for the greatest number of people (Balkan-sky, 1998, p. 54; Blitz, 1999, p. 583).

Such an index can be calculated through nearest-neighbor analysis, which measures the observedspacing between sites in a defined area, its relativedeviation from clustered or maximal dispersal, andchanges in these measures for a region over time(e.g., Earle, 1976). Greater articulation between dif-ferent tiers of the hierarchy, measured as mean dis-tances between sites, are also assumed to beindicative of a more integrated settlement system(e.g., Haggett, 1966). As Conolly and Lane (2006,pp. 164–166) note, however, nearest neighbor anal-ysis has several limitations. First, it was designed todetect spatial patterning of first nearest neighbors,and thus may not be sensitive to clustering at otherscales. The nearest neighbor index is also influencedby the size of the surrounding area included in theanalysis. As an alternate line of evidence therefore,relative degrees of integration and hierarchical rela-tionships between sites in the study region were alsoexamined through rank-size analysis.

Our rank-size analyses used total site area ratherthan mound count (including the revised values forLa Laguna) for comparative purposes, as this hasbeen the criterion used by most settlement analysts.Following conventional rank-size analyses practicedby geographers and archaeologists, the degree ofdeparture from log-normality in rank distributionsof settlement sizes has served as a useful measure inassessing the level of integration exhibited by partic-ular settlement systems (Haggett, 1966; Hodder andOrton, 1976; Johnson, 1977; Paynter, 1983; Zipf,1941, 1949; see concerns by Smith and Schreiber,2006, pp. 15–16 regarding the inclusion of non-urbansettlements). Recently Drennan and Peterson (2004)have proposed a valuable means of quantifying rela-tive departure from log-normality by summing thearea (A coefficient) over and under a log-normalregression line, measured as the shape of the devia-tion from a straight transverse line through a rectan-gular plane. An A value of zero denotes a log-normalpattern; positive or negative values can denote thearea of a convex pattern (positive), concave/primatepattern (negative), or the summed values of a primo-convex pattern (concave at upper left and convex atlower right) (Drennan and Peterson, 2004, p. 534).A values and error ranges employing Drennan andPeterson’s bootstrapping method were calculatedfor each PANT phase dataset through the RSBOOTprogram they have made available online (www.pitt.edu/~drennan/ranksize.html).

http://www.pitt.edu/~drennan/ranksize.html

http://www.pitt.edu/~drennan/ranksize.html

Table 1Summary of least cost path analysis (see Fig. 4)

Site Political apogee Least cost path through study region (yes/no)Estimated travel time in hours

GC 1 GC 2 GC 3 GC 4 MA

Cuicuilco Late formative No No No No No59.71 63.50 74.74 94.73 84.00

Teotihuacan Classic No Yes Yes Yes No48.89 57.26 69.78 89.96 82.19

Tula Early Postclassic No No Yes Yes No57.57 70.69 83.49 103.67 96.18

Texcoco Late Postclassic No Yes Yes Yes No51.85 58.24 70.36 90.54 81.12

Tenochtitlan Late Postclassic No No No No No57.24 63.46 74.77 94.77 84.97

Key: (GC 1) Gulf Coast point one, the northernmost, near El Tajin; (GC 2) Gulf Coast point two, 100 km southeast of GC 1; (GC 3) GulfCoast point three, 100 km southeast of GC 2, south of modern city of Veracruz; (GC 4) Gulf Coast point four, the southernmost, 100 kmsoutheast of GC 3, and near Matacapan; (MA) Monte Alban.


Four of the 78 sites from the Tlaxco phase wereomitted from the rank-size analysis because theylack size data. Three of these were listed as smallobservatories and the fourth is a small Postclassicreoccupation of the dominant Teotihuacan phasecenter (Merino Carrion, 1989, Cuadro 8); therefore,we do not believe that their omission significantlyalters our analyses or interpretations of the Tlaxcophase system. In addition to the modification ofLa Laguna’s recorded size, one site from both theTeotihuacan and Early Acopinalco phases wereassigned sizes based on estimated population and/or the ranking given by Merino Carrıon, becausethey lacked size data.3

General results

Relative value of the Tlaxcala Corridor

Our least cost path analysis supports GarcıaCook’s (1981) identification of the Tlaxcala Corridoras an important artery of communication for prehis-panic central Mexican civilizations, particularly dur-ing the Classic period. From Teotihuacan, thecorridor provides the easiest route to points east

3 The modification of size estimates from the charts in MerinoCarrion (1989) were as follows: Texoloc phase—T-491 (LaLaguna) from 30 to 66 ha; Tezoquipan phase—T-491 (LaLaguna) from 30 to 100 ha; Teotihuacan phase—T-164 fromnot recorded to 20 ha (based on possessing three ceremonialstructures and being classified as a ‘‘primary’’ center); EarlyAcopinalco phase—T-529 from not recorded to 1.5 ha (based onbeing classified as ‘‘rural’’ with an estimated population of 60);Tlaxco phase—omitted T-85, T-509, T-686A, and T-710.

and southeast, running from central Veracruz tothe Maya region. The northernmost point on theGulf Coast (GC 1) is better accessed by a northeast-erly route, identified by Charlton (1978, 1991) forits importance to Classic period obsidian exchange,and its continued usage in the Postclassic and Colo-nial periods. The remaining Teotihuacan–Gulf Coastroutes all run through the PANT survey region.Whereas GC 2 projects a route that makes partialuse of the Apulco Canyon, the route to GC 3 mirrorsthe modern Mexico City–Veracruz railroad and free-way, and GC 4, in the Tuxtla Mountains, is near tothe Teotihuacano enclave at Matacapan (see Santleyand Pool, 1993). The best route from Teotihuacan toMonte Alban begins in the Tlaxcala Corridor butbranches south before the study region.

Least cost path analyses originating from promi-nent central Mexican political capitals during otherperiods are summarized in Table 1. An interestingpoint of contrast is the utility of the corridor for tra-vel east and south from the southern Basin of Mex-ico (Cuicuilco and Tenochtitlan) versus from thenorthern Basin (Teotihuacan, Texcoco, and, to thenorth, Tula). Whereas the least cost path originatingfrom the southern Basin follows the route of themodern toll highway between Mexico City andVeracruz, traversing the high terrain between Iztac-cihuatl and Cerro Tlaloc into southern Puebla,those that originate from the northern Basin skirtthis range to the north and continue through Tlax-cala, often passing through the study region.

Our study does not consider improved transpor-tation capabilities using canoes within the Basin ofMexico lake system (Drennan, 1984; Hassig, 1985,


pp. 60–66) because we are interested in muchlonger-distance travel outside the Basin, andbecause it would involve the addition of multipleassumptions that move us too far away from theprimary dataset. As noted earlier, we also did notconsider altitudinal factors such as vegetative coverand weather. However, the addition of consider-ations modeling these factors would only serve toincrease the value of the relatively flat Tlaxcala Cor-ridor, particularly compared with the high-slopetravel of the southerly Basin–Puebla route. Withthese caveats in mind, our analyses provide strongevidence that the differences in the relative utilityof these alternate transportation corridors servedto structure the expansionary strategies of centralMexican polities differently, Teotihuacan and theAztec Triple Alliance in particular, together withthe sociopolitical geography of the time, as is dis-cussed as part of the social implications of theanalyses.

Site phases and continuity

Continuity in occupation of sites betweensequential phases provides a general indication ofhow disruptive certain periods in the history of thestudy region were. A limitation of the PANT data-set, however, is the irregular length of its occupationphases. While the Acopinalco phase is sub-dividedinto two very manageable centuries, the Tlaxcophase spans an unwieldy 600–650 years. Althoughproblems of determining contemporaneity of occu-pation affect all settlement surveys (Schacht, 1984),this attribute must be considered particularly wheninterpreting longer phases, which likely involvedfusion–fission and political cycling that are unde-tectable at this level of resolution (Anderson,1994; Blitz, 1999). Keeping irregular phase lengthin mind, the simple exercise of plotting the frequen-cies of sites with occupation spanning two succes-sive phases provides another line of evidencesuggestive of the impacts of macroregional develop-ments. By this measure, the Teotihuacan and LateAcopinalco phases appear to have been the mostdisruptive as they present sizeable decreases in sitecontinuity, with nearly half having been abandonedin both cases (Fig. 5a). We believe this pattern to besignificant irrespective of contemporaneity issues inboth cases because recent excavations confirm thatLa Laguna (the largest Tezoquipan center) wasabandoned during the Tezoquipan–Teotihuacanphase transition, and the Early to Late Acopinalco

transition bridges two short and well-defined chro-nological periods.

Site location and spacing

Nearest-neighbor values were calculated throughthe GIS for each phase to gauge the degree of spac-ing between sites in the study region, and their rela-tive tendencies towards clustering or dispersal. Thetwo most noteworthy patterns to emerge from theanalyses were (1) a general propensity of sites tocluster during all phases, and (2) greater dispersalof sites with ceremonial architecture (Tiers 1–3 ofour classification) relative to those without suchstructures in all phases except one (Fig. 5e). The firstpattern indicates communities in the region wereeither drawn to particular features of the landscapeor to other communities; the second supports theanalytical distinction we have drawn regarding thepresence/absence of ceremonial architecture, andsuggests that its presence is an index of higher-ordersocial functions (e.g., Smith, 1976), as is discussed inthe next section.

The tendency towards clustering lessens in theregion over time, but sites remain significantly clus-tered relative to their expected mean distribution(Euclidian distance) in all phases. The slight ten-dency to greater dispersal corresponds with a north-ward migration of the center of gravity (centroid) ofthe settlement region, calculated through the GIS asthe mean center of all sites (Fig. 5c). The northwardmigration of the settlement centroid corresponds toan increase in the mean elevation of sites and asharp spike in the occurrence of fortifications (Figs.5b and d). Whereas no sites are documented as for-tified or military observatories during the Formativeand Classic periods, approximately one quarter ofsites are during the Epiclassic and one-third of sitesare during the Postclassic.

Initial site clustering during the Formative phasesis associated with the southern portion of the studyregion, containing a higher proportion of sites at alower elevation along the Tlaxcala Corridor. The cen-ter of gravity of the settlements in the region movesnorthward in all periods until the Early Acopinalcophase, when it plateaus. The marked shift to moresites in the foothills of the northern mountain rangesbounding the Tlaxcala Corridor begins in the Tezo-quipan phase and continues during the Teotihuacanphase; however, the dominant centers during thesephases remain in the center of the corridor,unlike later phases when they are north of it. The

Fig. 4. Cost path analyses originating from five dominant central Mexican political centers, and depicting survey region (shaded), MonteAlban, and points along the Gulf Coast (GC 1–4 from north to south). (a) Teotihuacan; (b) Cuicuilco; (c) Tula; (d) Texcoco; (e) Tenochtitlan.


Fig. 5. (a) Frequency of sites that bridge the transition between two successive phases; (b) frequency of fortified sites; (c) latitude of theregional settlement centroid (mean center distance), showing the northward migration of sites; (d) mean elevation of all sites, and of siteswith ceremonial structures (Tiers 1–3); (e) nearest neighbor R values for sites with (Tiers 1–3) and without (Tier 4) ceremonial structures;(f) mean distances between sites and their closest higher-order center.


combination of these three patterns reveals a strongtendency for communities to become located in thefoothills of the Sierras of Tlaxco and San NicolasTerrenate, and to fortify themselves, particularly fol-lowing the collapse of the Teotihuacan system.

Site hierarchies and administrative functions

Potential hierarchical and administrative rela-tionships between sites were assessed through sev-eral independent lines of evidence with the

following results. (1) Nearest-neighbor analysesdemonstrate greater spacing between higher-ordersites (Tiers 1–3) relative to Tier 4 sites in all phasesexcept the Tezoquipan phase (Fig. 5e). (2) Meanspacing between sites and their closest higher-ordercenter suggests lower integration among communi-ties prior to the Teotihuacan phase, followed by apattern consistent with high integration during theTeotihuacan phase. The region witnesses a returnto lower integration during the Late Acopinalcophase, with greater integration once again in the


subsequent Tlaxco phase (Fig. 5f). (3) Rank-size dis-tributions suggest greater integration of the regionalsystem during the Tezoquipan, Teotihuacan, andEarly Acopinalco phases, and lower integration inthe phases preceding and following them (Fig. 6).

General consistency exists regarding the low indi-ces of inter-community and regional integrationduring the earliest phases (Tlatempa and Texoloc),the much higher indices of integration associated

Fig. 6. Rank-size hierarchies with A values and bootstrapped

with Teotihuacan and its initial period of decline(Teotihuacan and Early Acopinalco phases),marked ruralization during the Late Acopinalcophase, and the existence of multiple polities in theregion during the centuries prior to the SpanishConquest (Tlaxco phase). A more ambiguous issuederived from these data is the degree of integrationduring the Tezoquipan phase, just prior to theTeotihuacan phase. Measures of site spacing,

error ranges, following Drennan and Peterson (2004).


particularly between non-mound sites (Tier 4) andthose with ceremonial architecture (Tiers 1–3), aresuggestive of relatively low administrative integra-tion, while the rank-size distribution of the regionsuggests an integrated system with La Laguna asthe dominant center. These issues and others arediscussed in the following section, where we con-sider the social implications of the analyses.

Social implications of the analyses

Exchange and polity formation during the formative

period

Formative period settlement in the PANT regionwas closely associated with the Tlaxcala Corridor,with half of all sites and all ceremonial centerslocated within it. The earliest sedentary villageswere significantly clustered (Fig. 7), and the lowestR value (all sites R = 0.50; Tier 4 R = 0.52) coincid-ing with the Tlatempa phase is probably indicativeof early fissioning events resulting in groupings oflinked communities, as Earle (1976) documentedfor the eastern Basin of Mexico (see also Bandy,2004; Blitz, 1999). A two-tier hierarchy developedin the southwest of the survey region associated witha community possessing two ceremonial structures

Fig. 7. Tlatempa ph

(A = .251). Village autonomy may have been main-tained by all communities during the Tlatempaphase, or a small multi-village chiefdom may haveevolved including the Tier 3 site (T-505) and adja-cent Tier 4 sites.

The two primary settlement clusters of the Tlat-empa phase grew in the Texoloc phase, exhibitinga discernable shift in population eastward and thedevelopment of a Tier 2 site at La Laguna(Fig. 8). La Laguna was strategically situated atthe junction of the Tlaxcala Corridor and ApulcoCanyon. Its placement and significantly higher con-centration of ceremonial architecture suggest thatexchange and ritual functions were important fac-tors in its becoming the dominant Formative centerin the region, as was first suggested by Snow (1976,p. 10). Although spatial measures such as R valueand mean distance between higher- and lower-tiersites are suggestive of low levels of integration dur-ing the Texoloc (all sites R = 0.55; Tier 4 R = 0.58;Tiers 2–3 R = 0.83) and Tezoquipan (all sitesR = 0.58; Tier 4 R = 0.62; Tiers 2–3 R = 0.60)phases, the rank-size analysis is indicative of a sig-nificant increase in regional integration, with therank-size line changing from convex to primo-con-vex (Texoloc A = .249; Tezoquipan A = .105). TheTezoquipan A value is also the second closest to

ase settlement.

Fig. 8. Texoloc phase settlement.


zero of the seven phases, approaching log-normalityand suggestive of a slightly more integrated systemthan during the Texoloc phase.

The spatial analyses may be poor indices of com-munity integration and regional centralization dur-ing the Tezoquipan phase (Fig. 9) because all thehigher-tier centers (Tiers 2 and 3) were arranged ina strongly linear distribution running east–westthrough the Tlaxcala Corridor. This distributionaccounts for the lower degree of dispersal of theseceremonial centers, and the greater mean distancebetween them and Tier 4 sites. It also supports thenotion that trade through the Tlaxcala Corridorwas a major determinant in where larger communi-ties developed. As La Laguna sits near the south-eastern corner of the survey block, the spatialanalyses register it as being peripherally located inrelation to most Tier 3 sites—a boundary phenom-enon that requires survey data from adjacent areas,which is currently unavailable in a format that isreconcilable with the PANT database. Accordingto our analyses, the Tlaxcala Corridor would nothave served as the least cost path between the larg-est Formative central Mexican centers and the east-ern/southeastern Gulf Coast. Indeed the lineardistribution of the major Formative period sites—such as Cuicuilco, Tlapacoya, Tlalancaleca, and

Xochitecatl (see Fig. 2)—nicely fits the projectedleast-cost route running between Cuicuilco andthese sites.

In this light, the Tezoquipan settlement pattern isgenerally consistent with the political autonomy ofthe study region as a whole, and the formation ofsmall regional polities that benefited from theirpositioning along the corridor. The location of theTezoquipan settlement centroid in the center of anunoccupied region 15 km to the west of La Lagunamay be indicative of a buffer-zone between the pol-ity it controlled and one or more polities associatedwith the cluster of Tier 3 sites in the southwesternportion of the survey area. Although this boundaryline is speculative, its distance from La Lagunawould be consistent with a reasonable area ofadministration for a polity lacking state-levelbureaucratic administration, being located between12 and 16 km away, under a single day’s journeyby foot (e.g., Spencer, 1998). The nearest sites inthe southwestern cluster are within 25–30 km ofXochitecatl and Tlalancaleca, and may representthe northeastern limit of one of these polities.Again, due to the length of occupation phases wemust also envision the operation of a more fluid set-tlement system and allow for the possibility that fis-sioning processes and political cycling (e.g.,

Fig. 9. Tezoquipan phase settlement.

4 Surface mounds at T-85 were bulldozed in the 1970s, theirrubble being used in the construction of the dam for the ElCentenario reservoir. Therefore, there is no way of currentlyverifying the count of 40 ceremonial structures recorded byMerino Carrion (1989). However, visible surface artifact coversupports her 150 ha estimate.


Anderson, 1994; Love, 2002) affected the southwest-ern cluster of Tier 3 sites, as well as La Laguna,where excavations suggest a period of abandonmentor significant depopulation during the early Tezo-quipan phase.

Impacts of the urbanization and political expansion of

teotihuacan

A major reorganization of society in northernTlaxcala associated with the rise and expansion ofTeotihuacan are discernable through several linesof evidence: (1) the lowest level of settlement conti-nuity for all phases between the Tezoquipan andTeotihuacan phases (53% of sites); (2) the abandon-ment of the largest center (La Laguna) from the pre-ceding phase; (3) the establishment of a newdominant center that is the largest in size and num-ber of ceremonial structures for all phases; (4) theappearance of the only concave rank-size patternin the entire sequence (A = �.240), with boot-strapped error ranges indicative of the first signifi-cant departure in settlement system organization;and (5) the regular distribution of Tiers 1–3(R = 0.64) sites relative to Tier 4 sites (R = 0.55),suggestive of a nested hierarchy around the primary

center (all sites R = 0.55) (Fig. 10). The primarycenter—T-85, Cerritos de Guadalupe—was locateddirectly in line with the least cost path originatingfrom Teotihuacan to the southern Gulf Coast, andthe Maya region (GC 3 and GC 4). The site alsopossessed strong material affinities to Teotihuacan(Merino Carrion, 1989, p. 75), and was the only Tier1 site to develop in northern Tlaxcala during theentire sequence, following our classification scheme.Its 40 ceremonial structures greatly surpass thenumber recorded for the second largest site of thephase, with 7 mounds, as well as those recordedfor the largest sites during all other phases, with13.4 The combination of these lines of evidencestrongly support a dramatic restructuring of societyin northern Tlaxcala associated with Teotihuacanopolitical expansion and territorial administration.

Teotihuacano political expansion is generallyunderstood as having been selective and strategic,involving direct control of a relatively modest core

Fig. 10. Teotihuacan phase settlement.


area, more indirect control of the transportationcorridors leading from the core to other parts ofMesoamerica, and debated degrees of influence oversocieties outside of this core and corridors (e.g.,Cowgill, 2003; Hirth and Angulo Villasenor, 1981;Millon, 1988; Santley and Pool, 1993; Smith andMontiel, 2001). The Tlaxcala Corridor would haveserved as the most efficient transportation arteryconnecting Teotihuacan to large portions of theGulf Coast, Valley of Oaxaca, and Maya region;and during this period Garcıa Cook’s term Teotihu-acan Corridor truly applies. Classic period societiesin these regions possessed close connections to Teo-tihuacan, varying from economic and ideologicalexchanges to more asymmetrical political relations(e.g., Braswell, 2003; Marcus, 1983; Ruiz Gallutand Pascual Soto, 2004). It is likely that a majorityof these connections would have been channeledthrough the Tlaxcala Corridor.

Although sites in the Teotihuacan phase innorthern Tlaxcala do not become fortified, theirspatial segregation suggests that Teotihuacano pol-icy was to control movement through the corridorwithout concern for the direct administration ofthe entire region. The largest centers were locatedin the Tlaxcala Corridor, while many rural sites

were located at higher elevations in the foothills ofthe Sierra of Tlaxco. The relatively high degree ofsite continuity between the Teotihuacan and EarlyAcopinalco phases (80%) may also be indicative ofmore ephemeral Teotihuacano control of settle-ments north of the corridor, as the greatest continu-ity after the Teotihuacan phase is registered in thefoothills of the Sierra of Tlaxco. This area becomesthe center of regional developments in the EarlyAcopinalco phase (Fig. 11), while communitiesbecome scarce in the Tlaxcala Corridor, includingthe abandonment of T-85 and the reduction of thesecond largest Teotihuacan phase site (T-497) froma Tier 2 center to a Tier 4 hamlet.

Macroregional balkanization and the rise of

postclassic senorıos

By several measures, Early Acopinalco settle-ment presents the most regionally integrated phasein the PANT sequence. At first glance this may seeminconsistent with characterizations of the Epiclassicperiod of central Mexico as one of political region-alization or ‘‘balkanization’’ (e.g., Hirth, 2000, pp.244–269; Hodge, 1984, p. 13; Marcus, 1989), butis yet another example of how multiregional

Fig. 11. Early Acopinalco phase settlement.


phenomena and proper scales of analyses must beconsidered for settlement interpretations. Thehighly integrated settlement cluster that developedin northern Tlaxcala in association with the declineof Teotihuacan was safely buffered from the smallcity-states surrounding it on three sides: Teotihu-acan (still a city of approximately 40,000), Cacaxtla,and Cantona (see Cowgill, 2000; Garcıa Cook andMerino Carrion, 1998; Pina Chan, 1998). MerinoCarrion (1989, pp. 79–89) noted significant GulfCoast influences during the Early Acopinalco phase,possibly associated with the coastal site of El Tajın,or with Cacaxtla or Cantona, both closer highlandcenters with clear connections to Gulf Coast socie-ties. The Early Acopinalco settlement centroidreaches its northern latitudinal extreme, with a cor-responding spike in mean elevation of higher-ordercenters and the first appearance of fortifications,suggestive of increased hostilities. Rank-size andsite-spacing indices are consistent with a nested hier-archy of sites in the foothills of the Sierra of Tlaxcocentered on site T-606, San Miguel de las Piramides(A = �.061; all sites R = 0.67; Type 4 R = 0.67;Type 2–3 R = 0.78).

Although it is possible that Early Acopinalcocommunities were subordinate to a larger polity

such as Cacaxtla, Cantona, or Teotihuacan, it ismore likely that a small polity centered on T-606successfully retained its independence from centersoutside the study region. The sparse settlementwithin the Tlaxcala Corridor suggest that none ofthese polities directly administered it, and it is likelythat competition over the corridor on the part ofmultiple factions contributed significantly to thedestabilizing and eventual collapse of the Teotihu-acan system (Garcıa Cook and Merino Carrion,1996; Millon, 1988). Thus individuals of the EarlyAcopinalco polity likely acted as contributingagents of change in the macroregional politicaldynamics of the time, representing at the broadestscale one of the most significant ruptures in the his-tory of Mesoamerica.

At the scale of the PANT region only, the Earlyto Late Acopinalco phase transition is characterizedby widespread abandonment (only 55% of sites con-tinue) and ruralization (with only two mound cen-ters). A notable exception is the largest LateAcopinalco center (T-492, Tecoac, or Santa MarıaCapulac), which bridges the phase transition whileT-606 is abandoned (Fig. 12). Having the benefitsof relatively short Acopinalco sub-phases and exca-vation data from T-492, we are better positioned to

Fig. 12. Late Acopinalco phase settlement.


suggest that political cycling at the regional level canbe detected between the phases. Based on its cere-monial structures, T-492 was a Tier 3 center (withfive structures) while T-606 was a Tier 2 center (with13) during Early Acopinalco. In the Late Acopinal-co phase T-606 was unoccupied while T-492becomes a Tier 2 center in our classification by vir-tue of its eight structures (Merino Carrion, 1989,p. 90, mislabeled as domestic in Cuadro 7).

The spike in higher-order center nearest-neighborvalue (all sites R = 0.55; Tier 4 R = 0.55; Tiers 2–3R = 1.10) for the Late Acopinalco phase shouldnot be taken as indicating it was the most integratedsystem in the sequence, but is rather the result ofthere only having been two ceremonial centers at adistance from one another. Indeed mean-distanceand rank-size values characterize Late Acopinalcoas a period of regional decentralization. Most com-pelling in this regard is the rank-size hierarchy, withthe A coefficient moving to a strongly positive value(A = .378) and a bootstrapped error range near 80%confidence that it exceeds the mid-values of all priorphases. Whereas Early Acopinalco can be charac-terized as regional centralization during a periodof macroregional balkanization, Late Acopinalcocan be characterized as regionally balkanized.

The over 600 year duration of the PostclassicTlaxco phase makes settlement interpretations par-ticularly suspect for the issues of contemporaneitymentioned previously. Nearest-neighbor indicesare similar to those for the Teotihuacan phase (allsites R = 0.57; Tier 4 R = 0.59; Tiers 2–3R = 0.67). The rank-size distribution is stronglyconvex (A = .524), which could be indicative ofone of the following, non-exclusive possibilities:(1) the long duration of the phase obscures thesequential occupation of sites; (2) the area wasperipheral to the larger, imperial Aztec politicaleconomy; or (3) the convexity registers the poolingof multiple, generally contemporaneous systems(see Johnson, 1977, 1980; Paynter, 1983). Fortu-nately, some of the limitations of the lengthy Tlaxcophase can be partially offset by ethnohistoric datafrom the time of the Spanish Conquest.

Smith and Berdan (2003; Berdan, 2003) charac-terize Postclassic Puebla–Tlaxcala as a dispersed,yet actively competing core zone to the Aztec TripleAlliance imperial core centered in the Basin of Mex-ico. During the later Postclassic, however, indepen-dent Tlaxcala existed as a very different type ofcultural core; it was comprised of numerous smallNahua and Otomı polities who confederated to


resist incorporation by the Aztec Empire, which hadencircled them in an effort to deprive their inhabit-ants of desired resources such as salt and cotton.These hostilities may be registered by the fortifica-tion or elevated positioning of one-third of all sitesin northern Tlaxcala. Garcıa Cook and Merino Car-rion (1977, 1988, 1989) contended that three histor-ically documented Otomı centers of the region,which allied with the Tlaxcalan Confederation atthe time of the Conquest, can be archaeologicallytraced from the beginning of the Tlaxco phase.The three proposed Tlaxco phase centers are: (1)Atlancatepec in the northwest, (2) Tliliuhquitepecin the center, and (3) Tecoac in the southeast(Fig. 13) (Garcıa Cook and Merino Carrion, 1988,1989; Merino Carrion, 1989, pp. 93–125). All threemay have been the capitals of independent senorıos,or Atlancatepec may have been subordinate toanother polity (see also Anales de Cuauhtitlan

34:12, 45:11; Davies, 1968, pp. 66–155; Duran,1581/2002, p. 289, 383, 395, 399, 405, 599).

Based solely on public architecture the settlementdata does not support the existence of three inde-pendent polities, casting more doubt on the prehis-panic autonomy of Atlancatepec. One of thewestern Tier 3 sites would be the most likely candi-date for Atlancatepec, as they are near the modern

Fig. 13. Tlaxco pha

town of the same name. The two Tier 2 Tlaxcophase centers, however, correspond convincinglyas the centers of Tliliuhquitepec (T-630) and Tecoac(T-492, Santa Marıa Capulac, the largest LateAcopinalco phase as well; see also Trautmann,1984). The latter senorıo is of particular interestfor its potential as the point of entry into indepen-dent Tlaxcala by Cortes and his soldiers, passingthrough the Apulco Canyon from the town of Ixt-acamaxtitlan, and first battling an Otomı armybefore fighting and then forming an alliance withthe Tlaxcalteca against the Aztec Empire (Cortes,1519/1986b; Dıaz del Castillo, 1956, pp. 121–164;Munoz Camargo, 1584/2000, pp. 239–242). Westrongly reiterate Merino Carrion (1989, p. 98)assertion that historical Tecoac was definitivelynot the La Laguna site, as has been reported by sev-eral authors following Angulo Ramırez’s history ofTlaxcala from the middle of last century (AnguloRamırez, 2004, pp. 110–112). Merino Carrionrightly noted that small Postclassic settlements onthe hills surrounding La Laguna are spatially andchronologically disassociated with the Formativeceremonial center.

Considering that the ethnohistoric sources reportbetween two and three small Postclassic polities inthe region, settlement convexity is more likely

se settlement.


attributable to the pooling of contemporaneous, rel-atively peripheral systems, rather than chronologi-cally sequential occupations. The highest rankedcenters display a convex distribution within the90% confidence zone. Lower ranked sites follow adistribution that is roughly parallel to the expected,log-normal line, a characteristic more suggestive ofpooling than peripheral status (Paynter, 1983, p.259). Further, we consider the 99% bootstrappederror range for Tlaxco phase settlement having beenmore convex than Teotihuacan phase settlement asone of the most significant results of the study,and illustrative of the differing expansionary trajec-tories of Aztec Triple Alliance and Teotihuacan.

Prior to the arrival of the Spanish, the imperialAztec stranglehold rendered the Tlaxcala Corridorobsolete for interregional exchange. It was sparselyoccupied during the Tlaxco phase except for a clus-ter of ten sites in the southwest, recorded as havingcloser material affinities to the contemporaneousTexcalac and Tlaxcala phases of the confederatedTlaxcalan senorıos (Merino Carrion, 1989, Cuadro8). Aztec imperial strategy appears to have priori-tized securing easterly passage through the southernPuebla transportation corridor by conquering thisregion’s prosperous city states, such as Cholulaand Huejotzingo, over the Tlaxcala Corridor—which our least cost path analyses suggest was aslightly more valuable route from Texcoco but notfrom Tenochtitlan. It is likely that tributary andpolitical considerations were ultimately of greatersignificant in these decisions; although the relativecosts to the empire in subduing several small,loosely confederated, fortified polities bent on resist-ing incorporation must also be weighed (see Smith,1996). Settlement in the Apulco Canyon includedthree Tier 3 centers which may have participatedin interregional exchanges prior to Aztec imperialexpansion, but this corridor was likely of littleimportance to Late Postclassic economies (GarcıaMartınez, 2001).

Conclusion

Our multi-scalar reassessment of the PANT set-tlement data draws heavily on GIS and rank-sizedeterminations not available at the time of the sur-vey project in the 1970s; however, our social inter-pretations are largely complementary. Newinsights resulting from the study include our quanti-fication of the relative values of transportation cor-ridors to interregional exchanges passing between

central Mexico and points east and south, and theimportance of exchange to the development Forma-tive chiefly centers such as La Laguna, which wesuggest to have been the center of a polity compris-ing much of the eastern half of the study area.

Our reanalysis of the data is supportive of amodel of intensive administrative restructuring ofthe region related to Teotihuacan urbanizationand state expansion. Settlement changes associatedwith Teotihuacan include widespread abandon-ment, the establishment of a new regional hierarchycentered in the Tlaxcala Corridor at the largest com-munity in the sequence, and the movement of smal-ler sites into the foothills of the Sierra de Tlaxco.Although better illumination of the social processesassociated with Teotihuacano administration of theregion awaits further excavations, we propose theywere the strongest macroregional developments tohave affected northern Tlaxcala, implying relativelydirect control of this portion of Teotihuacan’s outerhinterland (see Hirth, 1978b, p. 331; Millon, 1988,pp. 113–114).

In stark contrast, Aztec imperial expansion didnot involve control of the Tlaxcala Corridor forthe following geographic and historic reasons: (1)the corridor was a relatively less valuable route fromTenochtitlan compared to Teotihuacan, whereas analternate route was more valuable; (2) this alternate,southerly route included rich city-states that mademore desirable tributary provinces than the smallpolities of northern Tlaxcala; and (3) the lack ofcentralization of the small polities in northern Tlax-cala, together with their active efforts at resistancethrough fortifications and elevated positions, likelyincreased the costs of conquering the region.

Methodological contributions of the studyinclude the use of GIS applications to derive addi-tional information from previously published data,particularly in terms of calculating transportationefficiencies of regional landscapes, and interpretingspatial relationships among sites, and betweensites and natural transportation corridors. Weadvocate that regional hierarchies based on cere-monial architecture serve a complementary func-tion to those based primarily or wholly on sitesize; the former has been common practice amongSoutheastern archaeologists, but infrequently usedby Mesoamericanists. By combining these twoindices as we have done, settlement analysts havericher lines of evidence at their disposal fromwhich to base their social interpretations of partic-ular datasets.


Multi-scalar approaches incorporating GIS offergreater resolution regarding transportation corri-dors, and regional dynamics in areas adjacent tocore areas of urbanism and political evolution.More rural regions such as northern Tlaxcalashould be considered peripheral to such develop-ments only in the sense that no major cities or statecapitals developed there. By taking a macroregionalperspective of settlement changes, however, we gaina greater appreciation for the importance of majortransportation corridors, the participation of theirinhabitants in the large-scale social transformationsof interest to archaeologists, and the broader con-text at which political evolution operated.

Acknowledgments

We thank Aleksander Borejsza, Jennifer Carb-allo, and four anonymous reviewers for their help-ful suggestions regarding the content andpresentation of the article. Any inaccuracies ormisrepresentations of the data are our own. Weare additionally grateful to John O’Shea and thejournal’s editorial staff for their assistance, andthe University of Oklahoma College of Arts andSciences for Faculty Enrichment Awards, whichallowed us to present the initial results of ourstudy at the 71st Annual Meeting of the Societyfor American Archaeology. Carballo thanks theNational Geographic Society and the Foundationfor the Advancement of Mesoamerican Studies,Inc., for their support of the Proyecto Arqueolog-ico La Laguna. We are appreciative of the pio-neering efforts of Angel Garcıa Cook andBeatriz Leonor Merino Carrion, who originallyassembled the data used in this study.

References

Aldenderfer, M., 1998. Montane Foragers: Asana and the South-Central Andean Archaic. University of Iowa Press, Iowa City.

Algaze, G., 1993. Expansionary dynamics of some early pristinestates. American Anthropologist 95, 304–333.

Anales de Cuauhtitlan, 1992. In: Bierhorst, J. (Ed. and Trans.)History and Mythology of the Aztecs: The Codex Chimalpo-poca. University of Arizona Press, Tucson, pp. 19–138.

Anderson, D.G., 1994. The Savannah River Chiefdoms: PoliticalChange in the Late Prehistoric Southeast. The University ofAlabama Press, Tuscaloosa.

Angulo Ramırez, A., 2004. Historia de Tlaxcala. Compiled andedited by Suarez Ruiz, R. TRLS Ediciones, Tlaxcala.

Balkansky, A.K., 1998. Urbanism and early state formation inthe Huamelulpan valley of southern Mexico. Latin AmericanAntiquity 9, 37–67.

Balkansky, A.K., 2002. The Sola Valley and the Monte AlbanState: A Study of Zapotec Imperial Expansion. Memoirs, no.36. Museum of Anthropology, University of Michigan, AnnArbor.

Balkansky, A.K., 2006. Surveys and Mesoamerican research: theemerging macroregional paradigm. Journal of ArchaeologicalResearch 14, 53–95.

Bandy, M.S., 2004. Fissioning, scalar stress, and social evolutionin early village societies. American Anthropologist 106, 322–333.

Berdan, F.F., 2003. Borders in the eastern Aztec Empire. In:Smith, M.E., Berdan, F.F. (Eds.), The Postclassic Mesoamer-ican World. University of Utah Press, Salt Lake City, pp. 73–77.

Blitz, J.H., 1999. Mississippian chiefdoms and the fusion–fissionprocess. American Antiquity 64, 577–592.

Braswell, G.E. (Ed.), 2003. The Maya and Teotihuacan: Rein-terpreting Early Classic Interaction. University of Texas,Austin.

Burghardt, A.F., 1971. A hypothesis about gateway cities. Annalsof the Association of American Geographers 61, 269–287.

Charlton, T.H., 1978. Teotihuacan, Tepeapulco, and obsidianexploitation. Science 200 (4347), 1227–1236.

Charlton, T.H., 1991. The influence and legacy of Teotihuacanon regional routes and Urban planning. In: Trombold, C.D.(Ed.), Ancient Road Networks and Settlement Hierarchies inthe New World. Cambridge University Press, Cambridge, pp.186–197.

Conolly, J., Lane, M., 2006. Geographical Information Systemsin Archaeology. Cambridge University Press, Cambridge.

Cortes, H., 1519/1986a. The First Letter. In: Pagden, A. (Ed. andTrans.), Hernan Cortes: Letters from Mexico. Yale UniversityPress, New Haven, pp. 3–46.

Cortes, H., 1519/1986b. The Second Letter. In: Pagden, A. (Ed.and Trans.), Hernan Cortes: Letters from Mexico. YaleUniversity Press, New Haven, pp. 47–159.

Cowgill, G.L., 2000. The central Mexican highlands from the riseof Teotihuacan to the Decline of Tula. In: Adams, R.E.W.,MacLeod, M.J. (Eds.), The Cambridge History of the NativePeoples of the Americas, Vol. II: Mesaomerica, Part 1.Cambridge University Press, Cambridge, pp. 250–317.

Cowgill, G.L., 2003. Teotihuacan and early classic interaction: aperspective from outside the Maya Region. In: Braswell, G.E.(Ed.), The Maya and Teotihuacan: Reinterpreting EarlyClassic Interaction. University of Texas Press, Austin, pp.315–335.

Davies, C.N., 1968. Los senorıos independientes del imperioazteca. Instituto Nacional de Antropologıa e Historia, Mex-ico City.

Dıaz del Castillo, B., 1956. The Discovery and Conquest ofMexico (Maudslay, A.P., Trans.). Farrar, Straus and Cudahy,New York, pp. 1517–1521.

Drennan, R.D., 1984. Long distance movement of goods in theMesoamerican Formative and Classic. American Antiquity49, 27–43.

Drennan, R.D., Peterson, C.E., 2004. Comparing archaeologicalsettlement systems with rank-size graphs: a measure of shapeand statistical confidence. Journal of Archaeological Science31, 533–549.

Duran, D., 1581/2002. Historia de las indias de Nueva Espana eislas de tierra firme, I. Consejo Nacional para la Cultura y lasArtes, Mexico City.


Earle, T.K., 1976. A nearest-neighbor analysis of two formativesettlement systems. In: Flannery, K.V. (Ed.), The EarlyMesoamerican Village. Academic Press, San Diego, pp. 196–223.

Falconer, S.E., Savage, S.H., 1995. Heartlands and hinterlands:alternative trajectories of Early Urbanization in Mesopotamiaand the southern Levant. American Antiquity 60, 37–58.

Flannery, K.V., 1999. Process and agency in early state forma-tion. Cambridge Archaeological Journal 9, 3–21.

Garcıa Cook, A., 1972. Investigaciones arqueologicas en elestado de Tlaxcala. Comunicaciones 6, 21–26 (FAIC, Puebla).

Garcıa Cook, A., 1981. The historical importance of Tlaxcala inthe cultural development of the central highlands. In: Sabloff,J.A. (Ed.), Supplement to the Handbook of Middle AmericanIndians Archaeology, vol. 1. University of Texas Press,Austin, pp. 244–276.

Garcıa Cook, A., Merino Carrion, B.L., 1977. Notas sobrecaminos y rutas de intercambio al este de la Cuenca deMexico. Comunicaciones 14, 71–82 (FAIC, Puebla).

Garcıa Cook, A., Merino Carrion, B.L., 1988. Integracion yconsolidacion de los senorıos en Tlaxcala, siglos IX al XVI.Arqueologıa 2, Instituto Nacional de Antropologıa e Historiaand Gobierno del Estado de Tlaxcala, Mexico City, pp. 155–168.

Garcıa Cook, A., Merino Carrion, B.L., 1989. La culturaTlaxco o senorıo de Tliliuhquitepec en el norte de Tlaxcala.In: Garcıa Moll, R., Garcıa Cook, A. (Eds.), Homenaje aRoman Pina Chan, Coleccion Cientıfica 187. InstitutoNacional de Antropologıa e Historia, Mexico, Mexico City,pp. 145–158.

Garcıa Cook, A., Merino Carrion, B.L., 1996. Situacion culturalen Tlaxcala durante el apogeo de Teotihuacan. In: Mastache,A.G., Parsons, J., Santley, R., Serra Puche, M.C. (Eds.),Arquelogıa Mesoamericana: Homenaje a William T. SandersI. Instituto Nacional de Antropologıa e Historia andArqueologıa Mexicana, Mexico City, pp. 281–316.

Garcıa Cook, A., Merino Carrion, B.L., 1998. Cantona: urbeprehispanica en el altiplano central de Mexico. Latin Amer-ican Antiquity 9, 191–216.

Garcıa Cook, A., del Carmen Trejo, E., 1977. Lo teotihuacano enTlaxcala. Comunicaciones 14, 57–70.

Garcıa Martınez, B., 2001. Los caminos prehispanicos y laestrategia de la conquista. Arqueologıa Mexicana 9, 44–47.

Gledhill, J., 1988. Introduction: the comparative analysis ofsocial and political transitions. In: Gledhill, J., Bender, B.,Larsen, M.T. (Eds.), State and Society: The Emergence andDevelopment of Social Hierarchy and Political Centraliza-tion. Unwin Hyman Press, London, pp. 1–29.

Haggett, P., 1966. Locational Analysis in Human Geography. St.Martin’s Press, New York.

Hassig, R., 1985. Trade, Tribute, and Transportation: TheSixteenth-Century Political Economy of the Valley of Mexico.University of Oklahoma Press, Norman.

Hirth, K.G., 1978a. Interregional exchange and the formation ofprehistoric gateway communities. American Antiquity 43, 35–45.

Hirth, K.G., 1978b. Teotihuacan regional population admin-istration in eastern Morelos. World Archaeology 9, 320–333.

Hirth, K.G., 2000Ancient Urbanism at Xochicalco: The Evolu-tion and Organization of a Pre-Hispanic Society, vol. 1. TheUniversity of Utah Press, Salt Lake City.

Hirth, K.G., Angulo Villasenor, J., 1981. Early state expansion incentral Mexico: Teotihuacan in Morelos. Journal of FieldArchaeology 8, 135–150.

Hodder, I., Orton, C., 1976. Spatial Analysis in Archaeology.Cambridge University Press, Cambridge.

Hodge, M.G., 1984. Aztec city-states memoirsMuseum ofAnthropology, vol. 18. University of Michigan, Ann Arbor.

Jennings, J., Craig, N., 2001. Politywide analysis and imperialpolitical economy: the relationship between valley politicalcomplexity and administrative centers in the Wari Empire ofthe Central Andes. Journal of Anthropological Archaeology20, 479–502.

Johnson, G.A., 1977. Aspects of regional analysis in archaeology.Annual Review of Anthropology 6, 479–508.

Johnson, G.A., 1980. Rank-size convexity and system integra-tion: a view from archaeology. Economic Geography 56, 234–247.

Kantner, J., 1997. Ancient roads, modern mapping: evaluatingChaco Anasazi roadways using GIS technology. Expedition39 (3), 49–62.

Litvak King, J., 1978. Central Mexico as part of the generalMesoamerican communications system. In: Lee, T., Nava-rette, C. (Eds.), Mesoamerican Communication Routes andCultural Contacts. New World Archaeological Foundation,Brigham Young University, Provo, pp. 115–122.

Love, M., 2002. Domination, resistance, and political cycling informative period pacific guatemala. In: O’Donovan, M. (Ed.),The Dynamics of Power. Occasional Paper No. 30, Center forArchaeological Investigations, Southern Illinois University,Carbondale, pp. 214–237.

Marcus, J., 1983. Teotihuacan visitors on Monte Alban monu-ments and murals. In: Flannery, K.V., Marcus, J. (Eds.), TheCloud People: Divergent Evolution of the Zapotec andMixtec Civilizations. Academic Press, New York, pp. 175–181.

Marcus, J., 1989. From centralized systems to city-states: possiblemodels for the Epiclassic. In: Diehl, R., Berlos, J.C. (Eds.),Mesoamerica after the Decline of Teotihuacan, A.D. 700-900.Dumbarton Oaks, Washington, DC, pp. 201–208.

Marcus, J., Flannery, K.V., 1996. Zapotec Civilization: HowUrban Society Evolved in Mexico’s Oaxaca Valley. Thamesand Hudson, New York.

Melville, E.G.K., 2000. Disease, ecology, and the environment.In: Meyer, M.C., Beezley, W.H. (Eds.), The OxfordHistory of Mexico. Oxford University Press, New York,pp. 13–243.

Merino Carrion, B.L., 1989. La Cultura Tlaxco. Serie Arqueolog-ica, Coleccion Cientıfica 174. Instituto Nacional de Ant-ropologıa e Historia, Mexico City.

Millon, R., 1988. The last years of Teotihuacan dominance. In:Yoffee, N., Cowgill, G.L. (Eds.), The Collapse of AncientStates and Civilizations. University of Arizona, Tucson, pp.102–164.

Munoz Camargo, D., 1584/2000. Descripcion de la ciudad yprovincia de Tlaxcala, second ed. Colegio de San Luis andGobierno del Estado de Tlaxcala, San Luis Potosı.

Paynter, R.W., 1983. Expanding the scope of settlementanalysis. In: Moore, J.A., Keene, A.S. (Eds.), Archaeolog-ical Hammers and Theories. Academic Press, New York,pp. 33–275.

Pina Chan, R., 1998. Cacaxtla: fuentes historicas y pinturas.Fondo de Cultura Economica, Mexico City.


Ruiz Gallut, M.E., Pascual Soto, A. (Eds.), 2004. La costa delGolfo en tiempos teotihuacanos: propuestas y perspectivas.Instituto Nacional de Antropologıa e Historia, Mexico City.

Sanders, W.T., Parsons, J.R., Santley, R.S., 1979. The Basin ofMexico: Ecological Processes in the Evolution of a Civiliza-tion. Academic Press, New York.

Santley, R.S., Pool, C.A., 1993. Prehispanic exchange relation-ships among Central Mexico, the valley of Oaxaca, and theGulf Coast of Mexico. In: Ericson, J.E., Baugh, T.G. (Eds.),The American Southwest and Mesoamerica: Systems ofPrehistoric Exchange. Plenum Press, New York, pp. 79–211.

Schacht, R.M., 1984. The contemporaneity problem. AmericanAntiquity 49, 678–695.

Schreiber, K.J., 1987. Conquest and consolidation: a comparisonof the Wari and Inca occupations of a highland PeruvianValley. American Antiquity 52, 266–284.

Sherratt, A., 2000. Envisioning global change: a long-termperspective. In: Denemark, R.A., Friedman, J., Gills, B.K.,Modelski, G. (Eds.), World System History: The SocialScience of Long-Term Change. Routledge, London, pp. 115–132.

Smith, C.A., 1976. Analyzing regional social systems. In: Smith,C.A. (Ed.), Regional Analysis: Volume II, Social Systems.Academic Press, New York, pp. 3–20.

Smith, M.E., 1996. The strategic provinces. In: Berdan, F.F.,Blanton, R., Boone, E., Hodge, M.G., Smith, M.E., Umber-ger, E. (Eds.), Aztec Imperial Strategies. Dumbarton Oaks,Washington, DC, pp. 137–150.

Smith, M.E., Berdan, F.F., 2003. Spatial structure of theMesoamerican world system. In: Smith, M.E., Berdan, F.F.(Eds.), The Postclassic Mesoamerican World. The Universityof Utah Press, Salt LakeCity, pp. 21–31.

Smith, M., Hodge, M., 1994. An introduction to late postclassiceconomies and polities. In: Hodge, M., Smith, M. (Eds.),Economies and Polities in the Aztec Realm. Institute forMesoamerican Studies, State University of New York,Albany, pp. 1–42.

Smith, M.E., Montiel, L., 2001. The archaeological study ofEmpires and Imperialism in pre-hispanic Central Mexico.Journal of Anthropological Archaeology 20, 245–284.

Smith, M.E., Schreiber, K.J., 2006. New world states andempires: politics, religion, and urbanism. Journal of Archae-ological Research 14, 1–52.

Snow, D.R., 1966. A Seriation of Archaeological Collectionsfrom the Rio Zahuapan Drainage, Tlaxcala, Mexico. Ph.D.dissertation, University of Oregon. University MicrofilmsInternational, 66–12, 984, Ann Arbor, Michigan.

Snow, D.R., 1969. Ceramic sequence and settlement location inpre-hispanic Tlaxcala. American Antiquity 34, 131–145.

Snow, D.R., 1976. Prehistory of the valley of Tlaxcala. In:Crawford, M.H. (Ed.), The Tlaxcaltecans: Prehistory,Demography, Morphology and Genetics, vol. 7. Universityof Kansas Publications in Anthropology, Lawrence, pp. 9–12.

Spencer, C.S., 1990. On the tempo and mode of state formation:neoevolutionism reconsidered. Journal of AnthropologialArchaeology 9, 1–30.

Spencer, C.S., 1998. A mathematical model of primary stateformation. Cultural Dynamics 10, 5–20.

Spencer, C.S., Redmond, E.M., 2006. Resistance strategies andearly state formation in Oaxaca, Mexico. In: Elson, C.M.,Covey, R.A. (Eds.), Intermediate Elites in Pre-ColumbianStates and Empires. University of Arizona Press, Tucson, pp.21–43.

Stanish, C., 2003. Ancient Titicaca: The Evolution of ComplexSociety in Southern Peru and Northern Bolivia. University ofCalifornia, Berkeley.

Stein, G.J., 2002. From passive periphery to active agents:emerging perspectives in the archaeology of interregionalinteraction. American Anthropologist 104, 903–916.

Steponaitis, V.P., 1978. Location theory and complex chiefdoms:a Mississippian example. In: Smith, B.D. (Ed.), MississippianSettlement Patterns. Academic Press, New York, pp. 417–453.

Tilly, C., 1990. Coercion, Capital, and European States, AD990–1990. Basil Blackwell, Cambridge.

Tobler, W., 1993. Three Representations of Geographical Anal-ysis and Modeling. National Center for Geographic Infor-mation and Analysis, Technical Report 93-1.

Trautmann, W., 1984. The impact of Spanish conquest on thedevelopment of the cultural landscape in Tlaxcala, Mexico: areconstruction using models. In: Harvey, H.R., Prem, H.J.(Eds.), Explorations in Ethnohistory: Indians of CentralMexico in the Sixteenth Century. University of New MexicoPress, Albuquerque.

Vega Sosa, C., 1981. Comparaciones entre Los Teteles deOcotitla, Tlaxcala y Teotihuacan a traves de materialesceramicos. In: Rattray, E.C., King, J.L., Dıaz Oyarzabal, C.(Eds.), Interaccion cultural en Mexico Central. UniversidadNacional Autonoma de Mexico, Mexico City, pp. 43–52.

Wheatley, D., Gillings, M., 2002. Spatial Technology andArchaeology: The Archaeological Applications of GIS. Tay-lor and Francis, Inc, London.

Wright, H.T., 1984. Prestate political formations. In: Earle, T.K.(Ed.), On the Evolution of Complex Societies: Essays inHonor of Harry Hoijer. Undena, Malibu, pp. 43–77.

Wright, H.T., 1986. The evolution of civilizations. In: Meltzer,D.J., Fowler, D.D., Sabloff, J.A. (Eds.), American Archaeol-ogy Past and Future: A Celebration of the Society forAmerican Archaeology 1935–1985. Smithsonian InstitutionPress, Washington, DC, pp. 323–365.

Wright, H.T., 2006. Early state dynamics as political experiment.Journal of Anthropological Research 62, 305–319.

Zipf, G.K., 1941. National Unity and Disunity: The Nation as aBio-Social Organism. The Principia Press, Bloomington,Indiana.

Zipf, G.K., 1949. Human Behavior and the Principle of LeastEffort: An Introduction to Human Ecology. Addison-WesleyPress, Cambridge, Massachusetts.

transportation corridors and political evolution in...

Documents