A few decades prior to the Spanish conquest of what was to become New Spain, the Aztecs conquered the native peoples of the Soconusco region, an important cultural frontier that lies along the Pacific seaboard of present day southern Chiapas, Mexico and adjacent Guatemala. Although the fact of this Aztec invasion was never forgotten, Mesoamercanists know little about the preconquest indigenous Soconusco people during the fifteenth century. This seems to us paradoxical because the early peoples who lived in the same region and who were among the first in Mesoamerica to make the transition to settled village life (e.g., Clark 1994, Clark et al. 1987, Voorhies 1976, 1996) have been relatively well studied. Why is it that currently more is known about people's lifeways five millennia before the current era, than about their cultural descendants who were drawn into the Aztec Empire in approximately A.D. 1486?

The answer can be found in the recognition of two trends in the research of ancient Mesoamerica. One of these is the past proclivity of Mesoamericanists to concentrate archaeological research on sites that boast elaborate artistic traditions and large public buildings such as those built by the inhabitants of the city-states of the Classic period Maya area and elsewhere. Since monumental construction was not a regular practice of the people of the Soconusco during the last centuries before the Spanish invasion, their sites tended to be overlooked. The lack of monumental architecture has significant consequences for site recovery as well, a point to which we shall return in a later section of this chapter.

The other pertinent trend in Mesoamerican research has been a curious lack of interest by archaeologists about the time touched by the long reach of the documentary record, a field of inquiry often left to historians. Although this trend has now changed, there can be no doubt about its adverse impact on research in the Late Postclassic period (A.D. 1200-1520) in Mesoamerica generally.

For the Soconusco these two trends, coupled with difficult field conditions, conspired in such a way that little scholarly progress was made in understanding the prehis-panic lifeways of its people during the thirteenth through sixteenth centuries. We fully recognize that many scholars have evinced interest in the late prehispanic culture of the Soconusco region and more generally in the coastal corridor of Chiapas and Guatemala (e.g., Becerra 1985; Coe 1961; Culebro 1975; Drucker 1948; Dutton 1958; McBryde 1947; Navarrete 1978; Orellana 1995; Palacios 1928; Piña Chan 1967; Seler 1900; Shook 1965; Thompson 1948; Vivó Escoto 1942). Usually, however, this interest has been tangential to other research concerns. None of these scholars has focused exclusively on the last prehispanic society of the Soconusco.

This monograph is an initial, albeit modest, contribution that is intended to partially redress this situation. It is the outcome of various archaeological and ethnohistorical research projects that we have carried out in the western portion of the Soconusco over a span of over 20 years. In it we present information about the material culture and settlement patterns of the ancient people living in the western Soconusco during the Late Postclassic period. We also seek to place the people of the Late Postclassic Soconusco in a larger context, exploring the ways in which trends in this region reflected those that took place in Mesoamerica generally and identifying those aspects of Late Postclassic Soconusco society that were unique.

In this introductory chapter, we set the stage for the presentation of the archaeological data by first providing some necessary background information. We continue with a review of the written record for the Soconusco region during the Late Postclassic period. Finally, we provide information needed to evaluate the archaeological data presented in subsequent chapters. We explain how we collected these data and discuss several problems that we consider to be of particular importance in this study.

The Soconusco Region

The Soconusco is located within the strip of flat-lying land that lies along the western edge of Chiapas, Mexico and adjacent Guatemala (fig. 1.1). This corridor begins on the north at the Isthmus of Tehuantepec, a low-lying, narrow waist of the North American continent in Oaxaca, Mexico. It angles southeastward, toward Central America, widening from approximately 25 km in the vicinity of the town of Tonalá to nearly 50 km at its southeastern end. The inland side of the corridor is formed by the Sierra Madre de Chiapas, a steep-sided mountain range that extends 200 km from the Isthmus, gaining in height steadily until it reaches the Tacaná volcano (4,110 m) on the Guatemalan-Mexican border. The seaward side of the plain has extensive wetlands, consisting of coastal lagoons, estuaries, and fresh and saltwater marshes (Voorhies 1989b).

At the time of the Spanish conquest, this elongated territory was known as the Province of Soconusco and included both the coastal plain and the adjacent foothills of the coastal mountain range. Previously, it had gained fame among the Aztecs for its superior cacao and for other tropical products such as jaguar skins and brilliant bird feathers (Gasco and Voorhies 1989; Voorhies 1989a). During colonial times, cacao production was an important commercial enterprise that linked the native populations of the Soconusco into a global economy (Gasco 1987). For approximately three centuries, cacao trees in the Soconusco supplied precious chocolate to the royal courts of Europe (Coe and Coe 1996:183) as well as to the closer markets of Mexico City, Oaxaca, Puebla and Guatemala City.

The Soconusco has the double distinction of being the last and most distant southern province to be dominated by the Aztecs for tribute purposes, as described below, and one of the last territories to be united with the Republic of Mexico (in 1842). In later centuries the Province of Soconusco was divided for administrative purposes into a northwestern Tonalá District--which coincided with the arid northern half of the coastal corridor--and the southeastern Soconusco District--with its higher rainfall, deeper soils, and perennially flowing rivers and streams. The waterways in the Soconusco made it a difficult region to traverse in the rainy season (Ponce de León 1961; Seargeant 1952). A railway, subject to annual washouts, was constructed along the foothills early in the twentieth century, but an all-weather highway was not built until the mid-1960s. In spite of seasonal transportation difficulties, the Soconusco from Preclassic times onward served as the principal "corridor" leading from north to south. Today Tapachula, the southernmost city on the Pacific coast of Mexico, is considered the modern "Gateway to Central America." Modern Soconusco exports include fish and shrimp, cacao, cattle, cotton, coffee, soybeans and other oil seeds, bananas and other fruit. Coffee production is limited to the upper piedmont and the mountains (Helbig 1964), but all of the other commodities are produced on the coastal plain or procured from the littoral and nearshore environments.

A vivid translation from de la Peña (in Lowe et al. 1982:59-61) details the particular physiography and climatic re-gime of the Soconusco, outlining their changing characteristics from north to south. Another socioeconomic study of the Chiapas coast (Bassols Batalla 1974) distinguishes between this northwestern sector and that of the Soconusco proper to the southeast. Other valuable recent descriptions of the Soco-nusco coastal plain are provided by Céja (1985:7-17, figs. 1-8) and Pailles (1980:3-18, figs. 2-10). Pailles includes a detailed physiographic map of a portion of the Mapastepec-Pijijiapan coast, prepared by Eduardo Martínez E. For the southwest Guatemalan coast, which includes the southern tip of the old Soconusco, an in-valuable description of the cultural and historical geography is provided by McBryde (1947), whose observations predate the era of radical modern change.

The intricate network of canals that fringes the coast provides ideal conditions for small boat transport. Here a system of barrier beaches protects the intercoastal waterway from the brunt of the forces of the Pacific Ocean with its high surf, treacherous undertow, and seasonal tempests. This inland communication route for boat traffic is known to have been used at least since Early Colonial times until the present century, and undoubtedly had prehispanic importance as well. (For a first-hand description of this natural canal system and the trade that moved along it, see Navarrete 1978:80-81; 1998). A detailed economic-geographical study of the Pacific coast of Chiapas made in 1950 also includes a reference to this inland waterway as a valued and very old natural economic resource (de la Peña 1951, vol.1:35). However, the absence of natural, deep-water ports on the 250 km long coast of Chiapas has meant that in modern times the Chiapas coast has not been important in commercial shipping.

An overview of the prehispanic archaeology of the territory as it was known up to 1961 is available in the Handbook of Middle American Indians (Lowe and Mason 1965). The most famous archaeological centers include Classic period Iglesia Vieja and Horcones in the Tonalá District (Ferdon 1953; Navarrete 1976, 1986) and the Late Preclassic to Early Postclassic period site of Izapa at the extreme southeastern limit of the Soconusco (Lowe et al. 1982). Some investigation of the Postclassic occupation along the Suchiate River at the southern Mexican-Guatemalan border region of the Soconusco was reported by Dutton (1958), but in general, little has been published of the late archaeology of this area.

For the early written history of the Soconusco region there are a number of important sources. A very early (1574) Relación de la Provincia de Soconusco was translated from the Spanish by Michael D. Coe (1961: see Ponce de León 1961 and remarks below). An overview of the Soconusco history, language and ecology is provided by Lowe et al. (1982:7-15,42-52,54-73). These authors include a linguistic map of Chiapas and a foldout topographic map of the Pacific coast of Chiapas and Guatemala (Lowe et al. 1982:figs. 1.1, 4.12). They also include a discussion of cacao production, with a lengthy translation from the 1586 Colonial period description of the Soconusco and early cacao use, orchards, and trade by Ciudad Real. A more intensive study of the signal role of cacao in colonial Soconusco has been made by Gasco (1987). A recent volume on the nature of the ancient prehispanic and posthispanic economy in the Soconusco has been edited by Voorhies (1989b).

Early Written History

It must be noted that all of the written records that pertain to Late Postclassic Soconusco come from the perspective of the region's fifteenth-century conquerors, the K'iche' and the Aztecs. Unfortunately, no indigenous written tradition has survived from the area. Evidence from Early Colonial documents (García de Palacio 1927; Ciudad Real 1976), as well as other linguistic evidence (Kaufman 1964; Campbell 1988:305ff.), indicates that the native language of the Soconusco was Tapachultec, a Mixe-Zoquean language; however, the only written native language that survives from the Colonial period is Nahuatl (e.g., Anderson et al. 1976:190-195), the language of the region's last prehispanic conquerors.

Both K'iche' and Aztec documents about the Soconusco tend to focus principally on the issues of conquest and resource extraction, leaving other matters aside. Although this situation limits what can be learned from documents, the available data are particularly useful for the analysis of economic and political affairs, for these were precisely the issues that concerned the conquering K'iche' and Aztecs. To reconstruct economic and political organization, we focus on three kinds of documentary data. First, we review information provided in the K'iche' and Aztec records about conquests of the Soconusco. We then turn to a discussion about Soconusco towns mentioned in the early documents. Finally, we review the data regarding the extraction of tribute from Soconusco towns. Like imperialistic societies throughout history, the Aztecs primarily were interested in determining what local resources could be extracted from the region, and in providing the infrastructure necessary to facilitate their flow to the Basin of Mexico. This information is particularly useful for understanding pre-Aztec economic and political conditions.

K'iche' and Aztec Conquests of the Soconusco

The earliest known written reference to Soconusco towns (table 1.1) is a K'iche' document, the Títulos de la Casa Ixquin-Nehaib (Recinos 1957, hereafter referred to as the Títulos), possibly written sometime between A.D. 1550 and 1560 (Carmack 1973:33), but referring to earlier events. This document describes K'iche' conquests that took place late in the fifteenth century under Q'uik'ab (and possibly by his successors), who ruled between 1425-1475 and was responsible for expanding K'iche' hegemony to its greatest extent.

According to this account (Recinos 1957:79-81), at least three Soconusco towns were conquered by K'iche' forces (fig. 1.2), descending from their homeland in the Guatemalan highlands. These are Ayutecat, Mazatán, and Tapaltecat. These ancient towns correspond to present day Ayutla, Mazatán, and Tapachula, respectively, all located relatively close to the international border between Mexico and Guatemala. A fourth named conquered town, Naguatecat, probably was in the Soconusco as well, although Recinos believes it is the modern Nahualate, a town located on the Nahualate River in the Guatemalan Department of Suchitepéquez. It seems to us more likely, however, that Naguatecat is the modern Soconusco town of Naguatán, located in the Department of San Marcos in western Guatemala. Naguatán appears consistently in Soconusco town lists for much of the Colonial period (archival notes in Gasco's files) and thus has a demonstrated deep history.

K'iche' control of southeastern Soconusco was short-lived, however, as a number of Soconusco towns, including Mazatán and Ayutla, fell to Aztec forces under the reign of Ahuitzotl, A.D. 1486 to 1502 (Hassig 1988:217). Although the details of this conquest are disputed (see Voorhies 1989a), the Soconusco towns of Mapachtepec, Mazatlan, Huehuetlan, and Huiztlan are listed in the Codex Mendoza (Berdan and Anawalt 1992:II, 22-25) as conquests of Ahuitzotl. In the same document, Motecuhzoma Xocoytzin, Ahuitzotl's successor, is credited with the conquest of Xoconochco and Huiztlán, which apparently had to be reconquered.

Additional documents that describe the Aztec conquest of Soconusco towns specify these same towns, as well as others (table 1.1). These documents include the Anales de Tlatelolco (1948:17-18), the Anales de Cuauhtitlan (1945:67), Book 9 of the Florentine Codex (Sahagún 1959), Alvarado Tezozomoc (1944:370ff), and Durán (1967: II, 387ff). Curiously, the town of Xolotlán is frequently linked with other Soconusco towns in these documents, but it does not appear in Aztec tribute lists or in any colonial documents, so we are unable to even guess its location or its relationship to other Soconusco towns.

Both Durán and Alvarado Tezozomoc present a chronological ordering of events, but the sequence of towns conquered is quite odd. Both chroniclers report that Aztec troops began their conquest of the region at Mazatán, then moved to Xolotlán, and then to Ayutla (the order of conquest of these two towns is reversed in Alvarado Tezozomoc). Soconusco was the last town to be confronted, and its inhabitants surrendered rather than engage in battle. This sequence of events would have the Aztec troops begin their conquest in the eastern portion of the region, eventually arriving at Soco-nusco, the town farthest to the west (fig. 1.2). One wonders how the troops got to Mazatán in the east without encountering hostile towns earlier. Perhaps they marched to the coast via the highlands (see Köhler 1978) or traveled by boat along the Soconusco coast until making landfall in the Mazatán region, a route discussed above and further below.

Towns and their Archaeological Identifications

All of the towns appearing in conquest and tribute lists of the K'iche' or Aztecs that are discussed by us are located in the southeastern half of what became the colonial Province of Soconusco. Sixteenth-century visitors to the area noted that people living in towns from the Tiltepec River (near presentday Arriaga, Chiapas) to the Tilapa River (which separates the Guatemalan Departments of San Marcos and Retalhuleu) spoke the same language, a Mixe-Zoquean language called Tapachultec by modern linguists (Ciudad Real 1976: I, 183ff.; Kaufman 1964; Campbell 1988:305ff.). This linguistic affinity may have been what prompted the Spaniards to include the entire area in the colonial Province of Soconusco.

We know very little about Late Postclassic occupation of the northwestern half of the province, from Mapastepec to the Tiltepec River. The Spanish chronicler Remesal (1964:81) mentioned Alvarado's destruction of a place called Sacrificadero, near Tonalá. Ferdon (1953:3) suggested that El Paredón, a Late Postclassic archaeological site (Navarrete 1959; and this volume) located on the shores of a coastal lagoon, might be Sacrificadero. We discuss El Paredón in chap-ter 3, but we cannot be certain that Paredón is Sacrificadero. It seems unlikely, however, that the archaeological site known as Tonalá, or Iglesia Vieja (Ferdon 1953), is Sacrificadero. Ferdon concluded that this site was abandoned prior to the Postclassic period, and he thought that most construction occurred midway through the Classic period. In addition, Drucker (1948) collected ceramics from the site, which Pfeiffer (1980) later identified as dating to the Late Preclassic through Late Classic periods. Accordingly, the Tonalá site cannot be the place called Sacrificadero by Remesal because it was abandoned by Alvarado's time.

There are at least two other Late Postclassic sites in the general vicinity of the modern town of Tonalá. Navarrete (1959:6) reports a Late Postclassic site located on the outskirts of town, near the Rio Zanatenco, on the western side of the modern highway (not to be confused with the nearby, but much more ancient site of Tzutzuculi). His map of the site (1959:6) shows nine rectangular alignments of foundation stones that are arranged in two rows.

Another Late Postclassic site is near Puerto Arista on the seaward side of Laguna La Joya. Called Cabeza del Toro, this site has been ravaged by looters (see Lowe et al. 1982:fig. 4.12 for map location). Many effigy feet from tripodal vessels, jade mosaics, and metal objects were removed over the years by fish merchants and others. Some of these artifacts are now in private collections in Tonalá and Tuxtla Gutiérrez (G. Lowe, personal communication, 1996).

Although we also know rather little about most of the towns farther to the southeast, at least there is name continuity for many of them. Both Becerra (1985) and Culebro (1939) listed most of the older towns in the Soconusco and discussed the meanings of indigenous names and the known ruins in their vicinity. We begin our discussion with Mapastepec, the northwesternmost of the towns (fig. 1.2) listed on table 1.1, and then examine the evidence for towns located farther to the southeast. A colonial pictorial map showing many of these towns can be seen in Navarrete (1978:fig.16).

Mapastepeque

We assume that prehispanic Mapastepec is located in the vicinity of the modern community of the same name, but the actual archaeological site has not been located.

Soconusco

It seems certain that the ancient town of Soco-nusco was located somewhere to the north of Acacoyagua, based on colonial descriptions of its location and on local oral tradition. We have identified two Late Postclassic sites in this area, which we refer to here by their local names, Las Gradas and Soconusco Viejo. In fact, the two sites may actually be different components of a single large site, but until the intervening area between the two sites is surveyed and tested, this conclusion remains tentative.

Las Gradas is located on a ridge top just to the west of the Cacaluta River. Voorhies and Gasco identified and briefly visited the site in 1998. It consists of extensive stone-lined terraces that extend up the north and south faces of the ridge. One large, stone-faced structure (approximately 12 x 12 m) and several other mounds are aligned along the east-west axis of the ridge, and there are heavy deposits of Late Postclassic sherds along the portions of the ridge top that we surveyed. There are also several large granite slabs, some standing upright, that may be plain stelae.

Soconusco Viejo is located a short distance to the east of Las Gradas on a bluff just east of the Cacaluta River. This is the site of the colonial town of Soconusco (Gasco 1990), and extensive Late Postclassic deposits extend over one kilometer to the northeast of the colonial town as well as to the west and northwest. The Late Postclassic component of the site was tested in 1997 by Gasco. The site consists of fifteen low earthen mounds. Household debris was found in virtually all of the test pits, whether or not they were close to mounds, indicating that many houses were located at ground level and suggesting that there was a large residential population at the site.

Las Gradas and Soconusco Viejo are almost certainly contemporary, but the relationship between the two sites is unclear; as mentioned, they may be two components of a single community. Under Aztec rule the town of Soconusco was the provincial capital, and Aztec records note that Aztec military officials were stationed there. Prior to our 1998 discovery of Las Gradas, we had tentatively identified the site of Soconusco Viejo as Postclassic Soconusco, yet we were not completely satisfied with this interpretation. This dissatisfaction stemmed, in part, from confusion in the archaeological literature regarding the location and description of the site. We also were troubled that Soco-nusco Viejo is such an unimpressive site. Why would such an apparently small, unimpressive town be designated the capital of the province, first by the Aztecs and later by the Spaniards? On the other hand, if Soconusco Viejo, Las Gradas and the area between them were part of a single, large community, such a designation would be easier to understand.

The identification of Postclassic Soconusco has been complicated by contradictory information in the literature. Drucker (1948:163) is the first to refer to the site in the archaeological literature. He mentions a site "claimed locally to be that of Soconoshco [sic]" and notes that it is "rather small and unimpressive." He did not investigate the site, it is not plotted on his published map (Drucker 1948:152), and he makes no mention of it in his field notes (Drucker 1947).

Navarrete (1978:77), who like Drucker conducted an archaeological survey in the area, says that he has identified the remains of Xoconusco "near the present town of Acacoyagua." On his unpublished regional map, the site appears a few kilometers northwest of that town, but this location could be either Soconusco Viejo or Las Gradas (archives NWAF, San Cristóbal de las Casas, Mexico)

Culebro (1975:14), a local historian, reports that the site is located northwest of the Hacienda Jalapa (a modern ranch close to Acacoyagua). He describes the site as consisting of pyramids that lie along an east-west axis and notes that several plain stelae are present. According to Culebro, the site occupies a large area in the foothills, and it has a series of steps built into the hillside leading to a shrine.

Although Drucker, Culebro, and Navarrete all claim to be describing the same site, it now seems to us that Drucker may be describing Soconusco Viejo while Culebro appears to be describing Las Gradas. We are uncertain which site Navarrete visited.

Gasco plans to conduct additional excavations at both Las Gradas and Soconusco Viejo (and the intervening area) in the near future, which should help resolve these issues.

Acapetahua

The site of Acapetahua is the only Late Postclassic center in the entire Soconusco region that has been investigated archaeologically. The prehispanic site lies adjacent to the modern town of Acapetahua, which presumably lies in the same location as the colonial town. In this case, the Late Postclassic population was moved a very short distance to reside in the colonial town. In chapter 2, we discuss the archaeological investigations at this site.

Huixtla

We have no information regarding the prehispanic location of Huixtla, but presumably it was located in the vicinity of the modern town.

Huehuetán

Drucker (1947, 1948) reports a Late Postclassic site, which he named Huehuetán I, three kilometers south of the Huehuetán railroad station, and he plots it lying to the east of the Rio Huehuetán. This site is discussed more fully in chapter 3. It seems likely that this was the site of Late Postclassic Huehuetán. Its location is approximately six kilometers from the public center of colonial and modern Huehuetán. This would mean that the native population was moved a few kilometers to occupy the colonial townsite, and this contrasts with the pattern we have seen at Acapetahua and the Colonial period site of Ocelocalco (Gasco 1987).

Mazatán

John Clark (personal communication), who has been working in the vicinity of Mazatán, reports that construction projects within the present-day town regularly turn up Late Post-classic artifacts. On this basis we hazard a guess that the preconquest town is at the same location as the modern one. Lowe (personal communication) notes that "Mexican-like" sherds have been recovered in the nearby Coatán riverbed and at Aguacate, on the east side of this river.

Tapachula

We have no information about the prehispanic location of Tapachula but assume it was located in the vicinity of the modern city of the same name. Lowe (personal communication) advises us that diagnostic Late Postclassic sherds have been recovered from the downtown area of the modern city.

Coyoacan

Coyoacan appears in colonial documents until 1573, after which it apparently ceased to exist. We have virtually no other documentary information regarding its whereabouts. We assume that it was located somewhere in the Cahuacan Valley because Cahuacan is a variant spelling of Coyoacan. Drucker (1948) located a Late Postclassic site, which he called El Chical, on the eastern banks of the Cahuacan River, a short distance from the coast (see chapter 3). Aztec-style sculptures from this site have been reported to us, and there is a large collection of Late Postclassic sherds from the site in the Museo Regional de Tapachula (INAH). Thus, it seems very likely that this site is prehispanic Coyoacan; unfortunately, the destruction of the site appears imminent, so future work may never be possible.

Tuxtla

We have no information about an archaeological site that may be prehispanic Tuxtla, or Tusta, but assume that it was in the vicinity of the modern town of Tuxtla Chico.

Ayutla

Archaeologist Michael Love has located a Late Postclassic site near present-day Ayutla (personal communication, 1990), on the Guatemalan side of the Rio Suchiate. In 1990 Gasco examined pottery collected by Love from the site's surface. Much of this material is similar to the Late Postclassic utilitarian pottery found in the Proyecto Soconusco study area. Other sherds are from types unfamiliar to us. This site is a likely candidate for Late Postclassic Ayutla.

Naguatán

We discussed earlier why the town of Naguatán may be the descendant of the prehispanic town with a similar name. There has been little archaeological work in this area of Guatemala, however, and we are unaware of any archaeological site that might be linked to the early documents.

Xolotlán

As noted above, Xolotlán is consistently mentioned with other Soconusco towns in the Aztec conquest lists. It does not appear in any colonial documents, however, and there are no known toponyms with a similar name in the region. As a result, we have no idea where this town was located.

The identification of these towns, and their tentative geographic placement, provides a political geography of native communities in the time period immediately preceding and just after the Spanish conquest of the area. Nevertheless, additional work must be carried out to definitively identify the towns discussed here.

Resource Extraction under K'iche' and Aztec Domination

We know few details of the ways in which the K'iche' obtained resources from conquered Soconusco towns. The Títulos specify that the following goods were taken from these towns by K'iche' forces, as booty or in tribute: cacao, cotton, pearls, gold, emeralds (probably jade or other greenstone), fish, and mantas (cotton cloaks) (Recinos 1957:79-80). Presumably, the K'iche' conquest of the Soconusco towns led to the imposition of regular tribute demands, but surviving documents are silent about this matter.

We have a much better understanding of the Aztec tribute system, and the manner in which resources were extracted from the Soconusco under Aztec domination. Elsewhere, we (Gasco and Voorhies 1989) have discussed the tribute from the Soconusco to the Aztecs (see also Hassig 1985), so we only briefly review this issue here.

Once towns in the Soconusco region had been conquered by the Aztecs, they began to pay tribute to the Triple Alliance. In two documents that record Aztec tribute records, the Codex Mendoza and the Matrícula de Tributos, eight Soconusco towns are named as tributaries. From northwest to southeast, these towns are Mapastepec, Soconusco, Acapetahua, Huixtla, Huehuetán, Mazatán, Coyoacan, and Ayutla (fig. 1.2). It is now thought that the tribute assessment recorded in these documents pertains to the years between A.D. 1516 and 1518 (Berdan 1992:65). Thus, they record the tribute paid by Soconusco towns as much as 30 years after Ahuitzotl brought the Soconusco region into the Aztec Empire. These documents, therefore, may reflect a situation in the region that was somewhat different from the one encountered when the Aztecs first conquered it.

A third Aztec tribute statement, the Información sobre los tributos que pagaban a Moctezuma año 1554 (hereafter called Información) (Scholes and Adams 1957), mentions only Soconusco and the existence of five other towns, which it fails to name. Reported tribute payments in the Información are much the same as those noted in the Matrícula de Tributos and the Codex Mendoza, but quantities differ.

Soconusco towns paid the following commodities annually in tribute (Gasco and Voorhies 1989):

• 2 strings of greenstone beads
• 4000 bunches of bird feathers
• 160 bird skins
• 2 amber lip plugs set in gold
• 2 blocks of amber
• 800 gourds (tecomates)
• 40 jaguar pelts
• 200 loads or cargas of cacao (1 carga = bundle of 24,000 cacao beans).

Implications for Late Postclassic Economic and Political Organization

In a general sense, what we glean from the documents is that in the Late Postclassic period the Soconusco region was rural and agrarian. There must have been substantial areas of wilderness that were habitats for the game whose products formed a major portion of the tribute. Nevertheless, there were also towns that were the seats of local government. Elsewhere we have discussed in some detail the nature of the Soconusco tribute payments to the Aztecs and what it implies about the nature of native society (Gasco and Voorhies 1989). Here, we expand a topic discussed by us in the 1989 article: the way in which the native peoples were organized politically and economically during the last several decades before the imposition of Spanish colonial rule.

Previously, Voorhies (1989a) presented a model for pre-Aztec political organization in which she proposed that in the Late Postclassic period the Soconusco region was made up of several independent polities, each with its own paramount center. In this model, the region later became unified, and the town of Soconusco became dominant only in the late fifteenth century after the Aztecs had conquered the area. Our objective here is to review the data for this and other scenarios and, in effect, to refine that model.

Details from K'iche' and Aztec documents that describe the conquest of and, in the case of the Aztecs, the subsequent collection of tribute from Soconusco towns allow us to construct a model of the political geography of the Soconusco in the last decades of the Late Postclassic period. An interesting detail in the Títulos lends support to Voorhies's argument that late prehistoric Soconusco was made up of independent polities. After the K'iche' took control of Mazatán, they enlisted the support of the Mazatecos to attack and conquer Tapaltecat (Recinos 1957:80). There is no hint that at the time Mazatán was subject to Tapaltecat, and, that in doing this, Mazatán was rebelling. Instead, the impression is that these were two independent polities and that the leaders of Mazatán were simply taking advantage of the situation, presumably with the hope that they could gain some favor with the powerful K'iche'.

The description of Aztec conquests in the Soconusco provides some useful information regarding the political relationships among these towns. Durán (1967:II, 387) states clearly that Mazatán was subject to Soconusco. Alvarado Tezozomoc (1944:374) reports that when Soconusco surrendered, its leaders lamented that "por nosotros ha muerto multitud de gente," which might also imply that Soconusco in some way dominated the other towns and thus felt responsible for the many deaths.

As we note above, Voorhies (1989a) has proposed that the eight towns named in the Matrícula and the Codex Mendoza may have served as administrative centers for eight independent polities prior to the Aztec conquest of the area. To us, this still seems the most likely explanation for the naming of these eight towns in these two documents. Other documents, however, suggest that at other times different towns may have served as administrative centers and that the number of such centers in the region did not remain constant. For example, the third tribute list mentioned above, the Información, named only the town of Soconusco but referred to another five unnamed towns, which suggests that at some time there were a total of six administrative centers in the region.

Presumably the K'iche', like the Aztecs, targeted administrative centers for conquest. This may mean that earlier in the fifteenth century Tapachula and possibly Naguatlán, along with Ayutla and Mazatán, were leading communities in the southeastern part of the region.

In 1530, only six years after the Spanish invasion, a somewhat different political configuration is suggested. At this time, six towns were named. Five of these towns were those named in both the Matrícula and Mendoza (Soconusco, Huehuetán, Ayutla, Mazatán, Coyoacan); the sixth was Tuxtla (today Tuxtla Chico). It is possible that these constitute the six towns mentioned in the Información, but we have no way to verify this.

In sum, in Late Postclassic Soconusco we seem to be witnessing a pattern in which several small polities existed, but the boundaries and relationships among them were constantly shifting. If we look at the documents that pertain to the last fifty years of the prehispanic era and the first six years of the Colonial period, at least ten (possibly twelve) Soconusco towns either were conquered by the K'iche' or Aztecs, or were tribute-paying entities, or both (table 1.1). Presumably, at one time or another, these towns were administrative centers or, as they would come to be called in the Colonial period, cabeceras. These towns were Mapastepec, Soconusco, Acapetahua, Huixtla, Huehuetán, Mazatán, Coyoacan, Ayutla, Tuxtla, and Tapachula, and possibly Naguatán and Xolotlán.

We know that many more than the ten to twelve towns named in K'iche' and Aztec conquest/tribute lists existed in the Soconusco in late prehistoric times. One early Spanish observer mentions that the province had at least 40 towns (CDIU 1842-1895:17:169). Presumably these unnamed Postclassic communities are the same ones that appear on Early Colonial town lists as anexos, or towns subject to the local administrative centers, the cabeceras. As we discuss more fully in chapter three, from archaeological data within one Postclassic polity, we speculate that there was a actually a three-tiered settlement hierarchy in place that consisted of primary centers (administrative centers or cabeceras), secondary sites, and tertiary sites.

Under Aztec domination, one town, Soconusco, was singled out as the provincial capital. Not only is Soconusco listed first on the tribute pages of the Matrícula and Mendoza, it is also the only town listed in the Información. Furthermore, in the Codex Mendoza, two high-ranking Aztec officials are named as "governors" serving in the town of Soconusco--a tezcacoacatl named Omequauh (Two Eagles) and a tlilancalqui possibly named Acueyotl (Wave).

We want to emphasize that the evidence is unequivocal that the town of Soconusco was the administrative center of the tributary province of Soconusco under the Aztecs. There have been reports in the literature (Drucker 1948:154; Thomas 1974:27; García Soto 1964:137; Vivó Escoto 1942:139; Campbell 1988:277) that Huehuetán was the capital of the province and that an Aztec garrison was located there. As far as we can tell, Vivó was the first to make this claim; Drucker, Thomas, and Campbell all cite Vivó Escoto, whereas García Soto does not provide any citation. Díaz del Castillo (1944:I, 368) is apparently the source for the notion that there was an Aztec garrison in the Soconusco region, but he states that it was in the town of Soconusco, not in Huehuetán. Others have questioned the interpretation of the term "garrison" as it has been translated from the Spanish (Davies 1978). But regardless of whether or not there was an Aztec garrison in Soconusco, Vivó Escoto is clearly mistaken when he claims that Huehuetán was the prehispanic capital. The source for this interpretation is a published 1565 document (Paso y Troncoso 1939-1942 Tomo X:63-65; see also Anderson et al. 1976:190-195). By this date, Huehuetán was the capital of the colonial Province of Soconusco. But this clearly was not the case during the period of Aztec domination.

In the period immediately following the Spanish conquest, Soconusco was the only town in the region that became an encomienda; it first belonged to Hernán Cortés but was given to Jorge de Alvarado in 1528. Typically, the towns that became encomiendas were city-state centers prior to the Spanish conquest (Berdan et al. 1996:109). The original cedula de encomienda, submitted by Alvarado in a court case, refers to "Soconusco y sus sujetos." Finally, in the 1530 tribute document mentioned above, Soconusco is explicitly referred to as the cabecera of the province. All of these factors make it clear that Soconusco became the dominant town in the region under the Aztecs, serving as the tribute collection center and the base for Aztec administrators. What is not at all clear is why Soconusco was chosen by the Aztecs to serve as the provincial capital.

The most obvious reason for the Aztecs to choose the town of Soconusco as an administrative center is that it was in some way higher ranking than other towns in the province. The accounts of Durán and Alvarado Tezozomoc, mentioned above, contain the only evidence to suggest that Soconusco might have played a dominant role in the region in pre-Aztec times. But, as Berdan (1996:123) notes, the towns that served as tribute collection centers in the Aztec Empire did not necessarily have political control over other polities in a tributary province. Perhaps Soconusco was made the administrative center for strategic reasons. Soconusco, situated in the western part of the province--closer to Tenochtitlan than all but one of the other conquered towns--may have been the safest location for tribute collection and a preferred point of departure for the porters who transported tribute items to Central Mexico.

Ideally, archaeological data might be used to clarify this issue, but currently we do not have site plans of all Late Postclassic centers in Soconusco that might allow us to evaluate their relative importance, nor do we know the precise locations of most of these sites, as we discussed above. We return to some of these issues in chapter 7.

The Field Studies

Much of the archaeological data that we report in the present monograph derive from a research project, directed by Voorhies, which she entitled Proyecto Soconusco. This project consisted of three field seasons: in the years 1978-79, 1981 and 1983.

During the first field season, in 1978-79, the team focused on locating archaeological sites within an 755 km² study area (fig. 1.3), mapping their surface features, and dating them, insofar as possible, by means of diagnostic potsherds recovered from their surfaces. Details of the survey methods are presented in Voorhies (1989c), so need not be reiterated here. Briefly, Voorhies relied on information from informants, combined with unbiased and scientifically rigorous site recovery methods to find sites. At the conclusion of the field season, 96 archeological sites had been found. The vast majority of the sites had surface features, typically mounds, that were mapped. Most sites could be dated by means of surface materials.

Thirty-one sites, or 32% of the inventory total, had Late Postclassic components. These are the sites that are discussed more fully in chapter 3. They form the bulk of the sites in the inventory used in our settlement pattern analysis, but sites found by other investigators are included as well.

The 1981 field season was dedicated primarily to mapping and digging exploratory test pits at the most imposing archaeological site that we had discovered in the previous field season. This site, Acapetahua, is the subject of chapter 2. During this second season, we identified the substantial Late Postclassic occupation at Acapetahua, situated along a prominent ridge top immediately adjacent to the Classic period public area on the flat coastal plain.

During this same field season, Hector Neff, then a graduate student at University of California, Santa Barbara (hereafter UCSB), conducted a study of the archaeological site of Las Morenas under the aegis of Proyecto Soconusco. Las Morenas is an intriguing site situated on an island within the coastal wetlands of the study area. Although Neff's study was severely limited by the lack of adequate resources, he was able to sketch the outlines of that site's history. A summary of these findings is given in chapter 3.

In 1983 Voorhies returned to Acapetahua in order to complete the test pit program and to excavate some architectural features. In addition, the crew excavated test pits at seven other sites: Lomas Juana, Rancho Alegre, Las Palmas, Ladrillo, Tepalcatenco, Filapa, and Las Lomas. We found Late Postclassic deposits at five sites, along with material from earlier and sometimes later time periods. The Las Palmas excavations, although limited to only two test pits, were particularly important because they produced pure Late Postclassic deposits. The Ladrillo materials were valuable for the same reason, although the diversity of cultural material was more limited than in the case of Las Palmas. Details of the excavations at these two sites are presented in chapter 3. In the case of Filapa, the only sherd scatter site without mounds in the inventory, Late Postclassic material was mixed with other material because of recent plowing. At Lomas Juana and Tepalcatenco, the Late Postclassic material was mixed with earlier material, but the reason for this situation was not immediately apparent.

Gasco's work at Ocelocalco also took place in 1983. This project, part of her doctoral work, was carried out independently of Proyecto Soconusco but was an outgrowth of her prior participation in the project. Gasco (1987, 1989a, 1989b) focused her work on the Colonial period architecture and associated materials at Ocelocalco, but she also collected information about the Late Postclassic occupation at the site and produced the first study that showed the pattern of material culture change from the Late Postclassic to the Colonial period (Gasco 1992).

Gasco's doctoral work had another major component that contributed to the present work. She conducted research at archives in Spain, Guatemala and Mexico and found rich information about the settlements during the Colonial period. These data are an important source of information that we have used in the reconstruction of the settlement pattern during the preceding Late Postclassic period.

Fieldwork in our study area did not end with the conclusion of Proyecto Soconusco and related projects. Voorhies returned to the field in 1988, at which time she focused on the reexcavation of one of the early shell mound sites in the coastal wetlands. Although her work that season did not produce information relevant to the Late Postclassic period, a coworker, Jack Mallory, conducted a survey along sections of the coast from the Proyecto Soconusco study area toward the Isthmus of Tehuantepec. Many of the archaeological sites that he visited, mapped, and surface collected proved to have had major Late Postclassic occupations. These sites, Paredón, Las Piedritas, Lomas del Camino, No Hay Lomas, Las Brujas, and Rancho Perú, are reported in chapter 3. Although Mallory conducted only a few test pit excavations, his descriptions of additional sites expanded our inventory from that period. Furthermore, several of the sites appear to have been occupied only during the Late Postclassic period. Because these sites were well preserved, they contributed significantly to our understanding of the intrasite settlement pattern for that time period.

In 1991 Voorhies conducted a second regional survey of the Soconusco coastal plain using a different strategy of site recovery, than that of her her first survey in 1978-79. The 1991 survey was restricted to an examination of riverbanks in order to find buried archaeological sites (Voorhies and Kennett 1995). Although Voorhies' objective was to find preceramic sites, those of other ages also were discovered in the course of the survey. The Huehuetán I and II sites, previously described by Drucker (1948), were visited during the 1991 field season and are reported in chapter 3.

The information gleaned from these various research projects forms the basis for this monograph. In addition to our own work, we have had access to information gathered by Drucker during a regional survey in 1947. We have consulted his field notes, which are now housed in the National Anthropological Archives, and have examined extensively the artifacts collected during that field research.

After we had completed the manuscript for this book, Gasco conducted a short season of fieldwork in 1997 at five of the Postclassic sites that are discussed briefly in chapter 3. The 1997 season consisted of excavating test pits at Soconusco, Ocelocalco, Las Piedritas, Las Brujas, and Lomas del Camino as well as mapping or modifying existing maps at Soconusco, Las Piedritas, and Las Brujas. A full discussion of this research awaits completion of the analyses of these materials and is not included here. In a few instances, however, Gasco's new data from the 1997 excavations is presented to clarify certain issues.

We must stress that the information reported here is not the result of a single problem-oriented project focused solely upon the subject at hand. Rather, what we have learned about the lifeways of people of the Late Postclassic period in this part of the Soconusco has accumulated over the years as we have pursued a variety of other archaeological objectives. Consequently, we present very basic information about the Late Postclassic period. Specifically, we document the kinds of tools and utensils that were utilized during that time span and make inferences about their former use. Second, we use the site plans to make inferences about the social organization of the community of people that once lived there and the arrangement of sites over the landscape as a means of reconstructing the regional organization as well. We are keenly aware that these contributions are limited in scope, but we are certain that they will greatly aid future researchers who wish to decipher the archaeological record left by the last prehispanic occupants of the Soconusco.

The Archaeological Record

There are several limiting aspects of the archaeological record that warrant discussion from the outset since they affect fundamentally the nature of our data: first is the relative lack of excavated material from single component occupations; second is the absence of a reliable chronology for the Soconusco Postclassic; and third is the difficulty of identifying sites with no surface features.

The Problem of Mixed Deposits

Most of the Late Postclassic artifacts that we report here derive from mixed deposits that contain artifacts of earlier time periods as well as Postclassic materials. At the beginning, we did not immediately recognize this situation because at first we were unfamiliar with the ceramic chronology of the region. This is not surprising because little prior work had been carried out in the western Soconusco, and the most important previous studies remained unpublished. Another contributing factor derived from our focus on the site of Acapetahua, where most of the excavations penetrated mixed fill. We now know that a great deal of land modification took place in ancient times at Acapetahua in order to widen and level the ridge top upon which Late Postclassic buildings were placed. As the project progressed, we eventually did excavate unmixed deposits of different time periods and learned to distinguish the basic ceramic chronology of the region. We now have great confidence in most of the tool and utensil types that we describe here, but the preponderance of mixed proveniences results in two consequences.

First, although detailed provenience data have been re-corded by us for all artifacts in the study, we have decided not to report these data when artifacts were recovered from mixed deposits because they are not likely to be useful for other investigators. Rather, we describe in a basic way the dating of sites, and, in the case of groups of artifacts, we present only the site and operation numbers, without giving level information. Our usual procedure for assigning a group of artifacts a Late Postclassic date was first to establish the presence of a member of the group in single component deposits that date to that time period. Once that was accomplished, we proceeded to describe the group of artifacts as a whole. In those instances where the artifacts evidently are not limited to one time period (for example, some morphological classes of metates), we often present a summary of the dating of the deposits in which a particular item was found.

Second, in a few instances, we are far from confident that a class of artifacts is actually Late Postclassic in age--specifically, where there are only a few objects in the group and all of them are from mixed deposits. We have indicated our uncertainty about chronological placement in the relevant discussions of artifact classes. We anticipate that future research will provide the appropriate correctives in instances where we have made mistakes.

The Problem of Postclassic Chronology

The chronology of the Postclassic period in the Soconusco is inadequately worked out at the present time. For instance, we were unable to formulate temporal phases based upon observed changes in artifact frequency over time because of the scarcity of stratigraphically unmixed deposits. More-over, we have so far been unable to identify satisfactorily all of the cultural materials dating to the Early Postclassic period (ca. A.D. 900-1200) in our investigations, apart from the hallmark Tohil pottery that Neff (1989: 11-5) identified at Las Morenas. This situation is not unique to our study area. For example, Henderson et al. (1979:190) note that in the Naco Valley of Honduras no material that dates securely to the Early Postclassic period has been recognized.

The puzzling apparent absence of an assemblage dating securely to the Early Postclassic period appears to us to be the consequence of insufficient research, resulting in a poorly defined chronology for the entire Postclassic period in the Soconusco. We seriously doubt that there was a precipitous decline in the regional population, an alternative explanation that cannot be rejected categorically on the basis of our current limited knowledge.

The possibility exists that the "missing" Early Postclassic material in our study area may actually be present but is attributed mistakenly to an earlier time span. Hector Neff (personal correspondence) reports that San Juan Plumbate (presumably Late Classic) and Tohil Plumbate (presumably Early Postclassic) both are found in association with local Late Classic pottery in Fred Bové's excavations in coastal Guatemala. Moreover, obsidian hydration dates associated with Late Classic pottery in Bové's excavations extend into the Early Postclassic period (A.D. 900 and 1200). Although the obsidian dates have yet to be verified by other chronometric analyses, Neff suspects that some "Late Classic" ceramic diagnostics may have persisted into the Early Postclassic period. Given the proximity of Bové's study area to our own, this is a plausible suggestion for the Soconusco as well. A similar situation has been noted by Zeitlin and Zeitlin in the Tehuantepec region, where certain ceramic types as well as sites that were once thought to date to the Late Classic period are now known to date to the Early Postclassic period (1990:427).

At Izapa, in eastern Soconusco, an Early Postclassic Remanso phase has been established and has been identified in 26 caches and dumps in the northern F Group (Lowe et al. 1982:153-157, 7.26; see also Lee 1978). Lowe et al. note that it is difficult to distinguish Late Classic utilitarian ceramics from those of the Early Postclassic (1982:157). The full Remanso phase ceramic complex has not been published, but it is based on the presence of Tohil Plumbate and large open vessels or basins that continue the Late Classic San Juan Plumbate tradition, as well as other non-Plumbate vessels including jars and tall cylinders.

Clearly, it is imperative that a refined chronology of the entire Postclassic period be established. Without a firm temporal framework, archaeologists cannot make solid inferences about the course of prehistory. Unfortunately, we are unable to resolve the problem of chronology on the basis of the data presented here, which has served only to identify the problem.

We emphasize, however, that the material discussed here constitutes a coherent assemblage that has been identified in unmixed, stratigraphically distinct deposits at the sites of Las Palmas, Las Morenas, Acapetahua, Ocelocalco, and Ladrillo. Therefore, there is no doubt that we are dealing with an archaeological component with internal integrity. We place this component within the Late Postclassic period on the basis of relative stratigraphic position and cultural cross ties with other assemblages in Mesoamerica.

A final note regarding chronological issues has to do with the traditional periodization of the Postclassic period in Mesoamerica. Whereas archaeologists in the past have tended to divide the Postclassic period into two parts, Early and Late Postclassic, there is a growing recognition that it may be preferable to create a tripartite division of Early, Middle, and Late Postclassic periods (see Smith and Berdan 2000). We suspect that many of the materials we describe here, which we assign to the traditional Late Postclassic period (ca. A.D. 1200-1520), will fit comfortably within a revised periodization scheme in which the Late Postclassic period is dated to ca. A.D. 1350-1520. Of course, this means that our previous discussion of the "missing" Early Postclassic may now become the missing Early (ca. A.D. 950-1150) and Middle Postclassic (ca. A.D. 1150-1350) periods. Overall, however, we agree that dividing the Postclassic period into three parts is warranted, and in future work we will attempt to identify the cultural characteristics of all three subperiods.

The Problem of Site Identification when Surface Features are Absent

On the south Pacific coast of Chiapas and Guatemala, we think that local people in the Late Postclassic period discontinued the earlier tendency to build platforms for their buildings. This change in building techniques has produced difficulties for modern archeologists, a problem encountered elsewhere in lowland Mesoamerica for this time period as well (e.g., Masson 1995). Previous surveys of the coastal plain of southwestern Guatemala led Coe and Flannery (1967:97, 99) to conclude that there was a "striking population decline" and the "near abandonment" of the coastal plain in the Postclassic period. Shook (1965), taking a more cautious position, noted that the Late Postclassic is poorly understood along the Pacific Coast, but he reported one Late Postclassic site that covered a level area of several acres "littered with cultural refuse" (Shook 1965:191). This site had "no mounds or slight undulations in the level terrain to indicate where houses of perishable material once stood," but he did find burned adobe fragments with impressions of poles, suggesting the presence of wattle-and-daub houses in the Late Postclassic period.

A decline in platform construction has particularly serious implications for any survey that relies on the presence of mounds for site identification. We have often been forced into just such a reliance because other strategies are not always viable within our study area.

In some areas, including the Mazatán region, the prevalence of plowing can reveal artifact concentrations in situations where no surface mounds exist. In the area reported here, however, plowing was not a standard agricultural practice at the time of our 1978-79 survey. In fact, in that survey we found only one sherd scatter site without mounds (Filapa), and, not surprisingly, it was detected in a freshly plowed field. Two years later, when we returned to excavate the site, it had disappeared from view altogether.

Given these experiences, we consider it extremely unlikely that sites with no mounds, or even with one or more very low mounds, can consistently be located using the survey methods that we used. It is possible, then, that Late Postclassic sites are severely underrepresented in our survey. In that case, any attempt to reconstruct the settlement hierarchy or demography for the Late Postclassic period on the basis of our data has the chance of being seriously compromised.

Despite this recognition, we have attempted to formulate a settlement hierarchy for the Late Postclassic period in the study area but have not ventured to reconstruct population size. In chapter 3, we present the settlement hierarchy, which follows our discussion, in chapter 2, of the primary Late Postclassic center in the study area. In chapter 4, we present the material culture we believe to be associated with food procurement and processing. In chapter 5, we focus upon tools for textile production, personal ornaments, sculpture, artisanal tools, and some miscellaneous items. In chapter 6, we describe ceramic vessels. Finally, in chapter 7, we summarize our findings and place them within a larger, pan-Mesoamerican perspective.