Archaeologists first reported Chau Hiix in 1989 (Pyburn 1991, 1992, 1998). The site lies slightly south of a line between Lamanai and Altun Ha, themselves less than 40 km apart (Figure 1), on the western shore of a seasonal lagoon. Chau Hiix contains about 25 large structures located on and around its central platform (Figure 2, Chau Hiix Site Center), the largest of which (Structure 1) is 20 m in height. The outlying settlement area is composed of a continuous scatter of platform groups and small mounds interrupted in at least eight places by clusters of much larger buildings on platforms along the western edge of Western Lagoon (Figure 3).

Waterways and archaeological sites in northern Belize.

The main platform at Chau Hiix, Belize.

Central and northern settlement areas of Chau Hiix.

Chau Hiix lies on a strip of “Hondo Silty” and “Xaibe” clays (Wright et al. 1959) between Western Lagoon and a relatively infertile pine savannah. The topography around the site ranges from flat to gently undulating, and the present agricultural potential of the local clays is limited by problems of drainage and transportation. Chau Hiix has immediate access to about 12 km² of well-drained prime agricultural land (Wright et al. 1959). Lamanai, neighboring to the west, was a much larger settlement with a larger and more fertile agricultural base (including some evidence for canalized fields, according to Alfred Siemens [1982]), while to the east, Altun Ha had a larger monumental center than Chau Hiix but the poorest land of the three communities. The “tropical moist” ecozone (Hartshorn et al. 1984) surrounding Chau Hiix consists of low broadleaf forest and cohune palm in drained areas. The lagoon and stream system south of Crooked Tree is famous for abundant birds, fish, and turtles that have provided subsistence and income for local Creole residents since the late 1700s. Evidence for the importance of these resources in prehistory is abundant, because most excavations produce turtle bone, fish (Deborah Vale, personal communication 2003), and freshwater clam and snail shells.

When occupied, Chau Hiix may have controlled portage between northern Belize and the Belize River Valley by an inland route between Yucatan and the central Peten. This location is especially advantageous for small boat traffic during the rainy season when hurricanes rage along the coast. Normally, water flows from Spanish Creek into Black Creek and into the Belize River, but for a few weeks each year, when floodwaters are high, the current is reversed, and Black Creek flows north from the Belize River. During this time, canoes can easily float down the Belize River from the Peten, then follow the current up Black Creek to Chau Hiix. From there, a portage of about 2 km to Dawson Creek would allow canoe access to the New River Lagoon north of Lamanai, then down the New River and out to sea without any upstream paddling. Local residents say this route was once possible without any overland portage, and it is likely that changes in land use have resulted in silt depositions blocking a connection to these passages from Western Lagoon to the New River system. Stream-flow reversal occurs at exactly the time of the year that oceangoing transport would have been affected by weather. Later in the wet season, when the creeks return to their normal flow patterns, this route is still a convenient and safe access for coastal traffic to inland areas. Chau Hiix occupies a crucial “choke point” on this route, which may have given its inhabitants access to trade and control over traffic flow.

Architectural and ceramic dates confirmed by David Pendergast and Robert Fry indicate that Chau Hiix was occupied continuously from the Early Preclassic (Swasey: 1100–900 b.c.) through the Late Postclassic (after a.d. 1500) period. To date, evidence for Swasey occupation is limited to redeposited pottery; Swasey sherds were discovered buried about 1 m deep within the platform at the base of Structure 1 mixed into a Terminal Classic (a.d. 800–900) deposit. The same pattern of redeposition was encountered at other locations on the central platform, a pattern suggested elsewhere to have ceremonial significance (Pendergast 1992; Walker 1990). Nevertheless, since all later periods (from 600 b.c. onward) are represented at Chau Hiix by primary deposits, very early occupation of the site area is conclusive.

Fluctuations in the water level of Western Lagoon obscure evidence of prehistoric land use, although some canals were found as early as the second season of exploration in 1991. Dry-season mapping has revealed that the entire lagoon system was once modified with an elaborate and extensive system of canals, dams, dikes, and wells (Figure 4). The limits of this system are currently unknown, but it appears to link several adjacent lagoons and wetlands and may include other dams and canals that have not yet been mapped.

Satellite image of Western Lagoon with key showing dams and canals.

A series of parallel ancient canals extending from east to west across the breadth of Western Lagoon becomes visible in the dry months of March and April. These features were originally about 5 m wide and 2–3 m deep (Figure 5). Excavation of lagoon canals shows their original depth and the depositional patterns that have developed since the abandonment of the canal system. The presence of a buried A-horizon soil below the lagoon canals and field areas indicates that prior to the Maya agricultural regime (which dramatically modified local hydrology), the patterns of seasonal flooding were different and perhaps less dramatic. This is suggested by the fact that heavy siltation began above what appears to have been a stable peaty surface that was probably at or near permanent ground water at the time. Depth to ground water is only about 1.5 m.

Dam excavation, Chau Hiix.

Ancient canals connect the site of Chau Hiix to an area of elevated land about 1 km away on the eastern side of Western Lagoon. Some of these canals (15 were visible in 1994) pass areas of ditching in the base of the lagoon that conform to the standard characteristics of canalized fields known from other parts of Belize (Harrison and Turner 1978; Pohl et al. 1990; Siemens 1982). Most of the largest canals have at least one deep well dug to the depth of ground water. These regularly spaced wells never completely dry up, even though they are heavily silted today. Predictably, silt deposits are deepest in the canals nearest the inflow of water from Spanish Creek, where the water slows considerably as it changes direction on entering Western Lagoon.

We suspect that these modifications were needed to control water flow because the onset, duration, and rainfall of each wet season is unpredictable in northern Belize. Two dams were built at the inlet from Spanish Creek into the lagoons around Chau Hiix that may have regulated the influx of water at the start of the wet season. Huge dikes (.5–.6 km in length) were constructed perpendicular to stream flow and parallel to the largest canals in Western Lagoon (Figure 4). Excavations into these dams indicate that they were built on the surface of the lagoon bottom with at least 1 m of sandy clay. These probably functioned to regulate the rate and timing of seasonal inundation and subsequent drainage of the lagoon and adjoining fields.

The canal wells could be used for pot irrigation during the dry season, making it possible to plant at the same time each year, despite variation in the onset of rains. Water fills the canals as the wet season begins in southern Belize before much rain falls locally and would have allowed irrigation if local rain was insufficient. As the wet season continues, southern rivers swell and back up, and Spanish Creek pushes water into Western Lagoon that would ultimately have inundated lagoon fields. Dams and dikes could regulate this to permit the late harvesting of low-lying fields. Flooding would then allow planting in the elevated areas along the seasonal lagoon's edge and on top of the broad dams. It remains to be discovered whether this system made it possible to plant more than one crop per year or simply allowed greater flexibility in responding to unpredictable weather and flooding. Annual inundation is essential for refertilization and killing tropical pests where monocropping is practiced, and riverine settings were important to early intensive agriculturalists worldwide. The system may also have played a role in aquaculture.

In addition to these enormous water works, which would have functioned to channel and organize excess water in the wet season, the opposite (eastern) side of Chau Hiix is modified by a tangle of terraces and reservoirs resembling those documented at the nearby settlement area of La Milpa (Tourtellot et al. 1999). Terraces and reservoirs, although as yet incompletely mapped, appear to be scattered along the entire western settlement area that borders the infertile pine ridge that runs parallel to the lagoon, bordering on the eastern edge of Chau Hiix's settlement. Although it is now impossible to calculate the productivity of this additional system, it is clear that it would considerably extend the agricultural potential of the site by increasing the amount and reliability of dry-season agriculture.

Ceramic dates from outside the site center suggest that settlement nearest lagoon fields was abandoned in the Terminal Classic period, while agricultural areas dependent on terracing and ditching continued somewhat longer into the Postclassic period. Agricultural fields themselves are difficult to date, but these residential changes may correspond to the onset of the development of the lagoon field system—in the Terminal Preclassic/Early Classic—and its abandonment in the Terminal Classic. An early Postclassic vessel was found smashed on the ground surface in the dry lagoon, above 20–30 cm of silt suggesting that Chau Hiix was occupied after the field system in the lagoon had been allowed to deteriorate. The timing of the introduction of a new and extremely productive system of agricultural exploitation into a relatively small community of smallholders with the rise of major centers of the Classic period is not coincidental. Nor is the abandonment of the Chau Hiix lagoon edge settlement nodes in the Terminal Classic simply coincidental with the depopulation of Altun Ha as well as more distant political centers. Neglect of the Chau Hiix canal system in the Terminal Classic would have both decreased the productivity of the local community and made it difficult to access.

Without continual maintenance of the canals, the arterial waterways would rapidly become impassible during most of the year. Today the increased vegetation on silted areas makes passages invisible even when there is sufficient water present for canoe traffic. In effect, the residents of Chau Hiix not only disengaged from a larger interaction sphere; they were able to hide from potential threats from outside (Andres and Pyburn 2003). At present, however, we are unable to ascertain whether the abandonment occurred before or after the arrival of the Spanish; we suspect earlier due to the apparent depopulation of lagoon edge sites before the Postclassic period, but as of yet we have no datable material culture associated with use of the canals.

The implications of this scenario for understanding Maya settlements are important. I have argued elsewhere (Pyburn 1998; Pyburn et al. 1998) that at least some Maya residential patterns relate more to long-distance relations of trade and exploitation than to local soil quality or potential for intensification. Clearly there are areas of the Maya Lowlands where some forms of intensification would have been possible but did not develop, while other areas, some with relatively modest population densities, appear to have intensified their production well beyond either local, or even regional, requirements. By common measures of carrying capacity, Chau Hiix appears to have had sufficient arable land to feed its residents, even in the Classic period when the settlement nodes along the canal were expanded, without resorting to the development of the lagoon fields, which would have channeled as much as 7 million m³ of water (Jerry Fairly, personal communication, 2000) and made as much as 10 km of land extremely fertile, possibly allowing more than one crop per year. No such modifications have been identified at either Lamanai or Altun Ha. It may be that Chau Hiix supplied these nearby sites with agricultural produce, but because land quality and availability do not suggest either was stressed by land shortages in their immediate area, the implication is that if either community depended on food from Chau Hiix, it was because they were supplying more than local needs. Again, the location of Chau Hiix on an arterial water route is surely not coincidental.

A report on the discovery of the ancient hydrological modifications around Chau Hiix belongs in a volume that pays tribute to Pat Culbert for two reasons. First of all, it clearly echoes a significant part of his contribution to our understanding of the Maya. Long before the idea was common knowledge, Culbert made it clear to his students that microenvironmental variation and local knowledge would profoundly affect ancient Maya subsistence strategies, but also that sweeping characterizations of Maya agriculture such as “swidden” or “raised field” would never suffice to characterize the Maya achievement. He insisted that we look closer and showed us how to do it.

But perhaps more important, this report belongs here in recognition of Culbert's lifelong commitment to the integration of perspectives. For him, archaeology has always been embedded within cultural anthropology, a means of expanding our models and comparing our data over wider reaches of time and space. And it is that emphasis on cultural context that led to his emphasis on the collaboration between epigraphy and dirt archaeology in the reconstruction of Maya culture history. These two lines of evidence must be used to rule out competing hypotheses, whether they pertain to dynastic history or stratigraphic continuity.

In conclusion, top-down models of Maya organization based on those hieroglyphic inscriptions that fortuitously have been preserved offer a picture of Maya political order in terms that were set by Maya politicians themselves. The idea that ancient Maya centers were all equal competitors in a game of segmentary state power without real economic consequences elides the reality of a political economy based on economic hegemony and exchange between cities. It is not hard to envision why this should be so: Maya rulers were in the business of reality construction. But it is important not to accept their claims to power at face value; their insistence that their authority came from their personalities, godlike omniscience, absolute conquest, and excellent heredity requires a skeptical eye. In short, it is important not to believe what they were so eager for their contemporaries to believe—that their power came from ideology—but to realize that their power also came from control of labor and the means of production. Belief in the rightness of rulership helps, but it does not substitute for hegemony and exploitation—it only makes domination easier (Scott 1992) and the inevitable resistance more subtle. Hieroglyphic interpretations once seemed to support the unilinear reconstructions of Maya political economy proposed by archaeologists working with incomplete knowledge of Maya environmental and production diversity and before Culbert insisted on a more complex picture. As more evidence of intersite hegemony appears, the time has come to rethink models based on cultures influenced by the modern world system and the implicit belief in the inevitability of a particular route to complex society and civilized life (Wolf 1982). In the end, Culbert's views have pushed Maya studies into a new era, away from the simplified concepts that once gave a scholarly impetus to the New Archaeology and into a more complex realm in which archaeological visibility and the evolution of political economy can no longer be empirically equated with a rise or a collapse.