ABSTRACT
Located in the northwest Yucatan Peninsula, Chunchucmil was a major, densely populated Maya center during the Late Classic Period. The high population density of this site remains a mystery in light of the apparently low agricultural potential of its setting. This area today is extremely dry and the surface is covered by vast outcroppings of limestone with thin and meager soil cover. Possible answers to the Chunchucmil paradox are: 1) Chunchucmil agriculturalists used agricultural intensification techniques to increase production, and 2) Chunchucmil was involved in regional trade of coastal resources (such as salt) for food. Understanding the settlement pattern, site boundaries, and subsistence base of Chunchucmil requires extensive mapping of the settlement, including a systematic soil survey. As a precursor to an extensive program of research, we conducted a preliminary soil and environmental survey in 1994. This paper presents the results of that analysis, considers alternative agricultural practices, and suggests the direction of future research at the site.

INTRODUCTION The modern village of Chunchucmil in the northwest Yucatan Peninsula of Mexico (Figure 1) is located 27 kilometers inland in a region in which climate (700 - 1000 mm. annual rainfall) and generally thin soil limit agricultural potential. However, during the Maya Late Classic Period (AD 600 - 900), Chunchucmil was a densely populated Maya center. This high population in a region of limited access to fresh water and productive soils has puzzled scholars working in the northern lowlands (Vlcek et al. 1978). The Regional Economy Project, directed by Bruce Dahlin of Howard University, is currently working on the Chunchucmil problem. As part of the project, we conducted a preliminary soil survey in 1994 to establish the context for assessing agricultural potential. This paper presents the results of that survey and its implications for understanding ancient settlement at Chunchucmil.

POPULATION ESTIMATES
Preliminary mapping work in the Chunchucmil region estimated a site center of 30 ha., a central residential area of 6 km², and an outer rural zone of 13 km² (Vlcek et al. 1978, Kurjack and Garza 1981). Vlcek estimated that the central residential area of Chunchucmil had a population of 12,000. This estimate is based on a minimum of four houses per hectare, a conservative number, with five persons per household. This figure is as large as the modem population in a 2,000 km² region surrounding the site. Vlcek estimated the population of a 13-km² outer rural zone at between 8,000 and 18,000. Areal estimates were based on remote sensing images of the site and field reconnaissance. Dahlin's subsequent field observations indicate that the site center is larger than Vlcek reported. Much of the site is obscured by secondary forest growth, and rigorous estimates of the areal extent and population of Chunchucmil await the intensive mapping proposed by the Regional Economy Project. Nonetheless, using even the most conservative estimates, Chunchucmil had a high[end p. 41] population density, like many Late Classic Maya sites.

AGRO-CLIMATE AND PAST CLIMATE
The physical geography of the area, namely the karst topography and dry climate, and the water requirements of maize, the presumed staple crop of the Chunchucmil Maya, raise questions about its production and the subsistence base of Chunchucmil. The rainfall pattern of the northwest Yucatan Peninnsula will not allow highly productive maize cultivation. Maize requires 500 mm. of rain during the growing season (Wellhausen 1957). This much precipitation is typical, but all of it is not available to plants. The Köppen climate designation places the region near the BS-Aw (steppe-tropical savanna) boundary (Wilson 1980); the area receives 700 - 1,000 mm. of rain annually (Carta de Precipitación Total Anual, 1981), 80 to 90% of which falls from July through September. Most of the moisture is lost by karstic seepage through the limestone or by evapotranspiration, and there is a mean annual water budget deficit of 600 to 700 mm. (Carta de Evapotranspiración y Deficit de Agua, 1983). Precipitation is also highly variable, with annual deparrtures from the long term average of as much as 30%. Variation is accentuated by a canicular drought and highly localized precipitation (Wilson 1980).

Maize production and consumption figures are available for Komchén, a village 76 km. northeast of Chunchucmil (Figure 1), with comparable soils and a drier climate (600 to 700 mm. annual rainfall) than that of Chunchucmil. According to estimates for Komchén, a 10 to 12 year fallow cycle would [end p. 42] require approximately 12.5 to 14.5 ha. per person (Shuman 1974). Using these figures and Vlcek's conservative size and population estimates, an area of 2,500 to 2,900 km² would have been required to feed Chunchucmil's population. It seems highly unlikely that the population of urban Chunchucmil could have supported itself by maize production. Circumstantial evidence of the difficulty of agriculture in this area is provided by Hacienda Chunchucmil, which was operating in the 19th and early 20th centuries. Originally established for commmercial maize production, it switched to cattle production due to low maize yields. Eventually, even cattle-raising failed, and the hacienda owners turned to henequen cultivation (Vlcek et al. 1978).

Climatic conditions in the Yucatán region during the Late Classic Period remain the subject of debate (Pohl 1990, Rice 1993), but growing conditions were probably not much better in the past. Sediment analysis indicates no significant soil loss, and we infer that the soil cover was as thin in antiquity as it is today (Dahlin et al. 1991, 1992; Whitmore et al. 1994; Leyden et al., in press). Paleoclimatic reeconstructions show an unstable vegetative regime, but one that was generally similar to that of today. Pollen and phytoliths in sediment cores from Cenote San José Chulchaca, 24 kilometers north of Chunchucmil (Figure 1), indicate that the forest was a similar low scrub; the soils were thin from the mid-Holocene to the present (Leyden et al., in press). The evidence suggests that Late Classic Period precipitation in the northern Yucatán was slightly greater than previously and was more seasonal in its distribution than it is today. Recent research with oxygen isotopes and geochemical indicators in a core from Lake Chichancanab in the central Yucatán peninsula (Figure 1) and San José Chulchaca (Hodell et al. 1995, Leyden et al., in press) has indicated the onset of drier conditions than today at the end of the Late Classic (ca. AD 800 -1000). This arid episode would have had a severe effect on agricultural potential.

ALTERNATIVE ECONOMIES
Given that climatic conditions were similar to (or drier than) today and the unlikelihood that maize production alone could have supported the population of Chunchucmil, we looked at alternative economies for the residents of Chunchucmil. Either food was imported via trade, or agriculture was intensified extraordinarily (by fertilizing, mulching, pot irrigation, or seed-bed cultivation).

The Chunchucmil site, occupying the ecotone between karstic plain and estuarine coast, is strategically located for regional trade. This locational advantage is strengthened by the site's proximity to the Puuc region (Figure 1), which likely produced a food surplus through the Late Classic Period (Dunnning 1992a). A prehispanic maritime trade route, inncluding Gulf Coast sites such as Isla Piedras, Uaymil, Jaina, and Punta Canbalam, provided an avenue of trade for highland and lowland sites in Mexico and Guatemala. This route included the northwest Yucatán sites of Chunchucmil, Oxkintok, Sayil, and Uxmal (Andrews 1990). The coastal site of Canbalam, 14 km. south of Celestún, was likely the major trading port for Chunchucmil.

Vlcek et al. have hypothesized that it was salt, not agricultural suitability, that determined the location of Chunchucmil (1978). The northwest coast of Yucatán is composed of beach ridges and swales which act as basins (salinas) for the collection of salt (Figure 2). These salinas have historically been important to the regional economy and remain so today. The Celestún salinas were the second largest salt flats in Mesoamerica (Andrews 1983; Roys 1957). It is reasonable to assume that they were important during the Late Classic Period as well, but salt production in the economy of ancient Chunchucmil has yet to be fully explored. If salt were an important commodity, Chunchucmil may have been a regional trading center for the exchange of coastal and inland goods and services. [end p. 43]

In recent decades, archaeological investigation in the lowlands has broadened its focus, shifting some of its emphasis from monumental architecture and elites to the economies of entire communities. These studies have shown that the Maya had diverse agricultural systems, including such intensive techniques as terracing, draining fields, gardening, mulching, and orchard cultivation (Beach and Dunnning 1995, Chase and Chase 1983, Dunning 1992b, Flannery 1982, Gómez-Pompa et al. 1987, Harrison and Turner 1978, Turner and Miksicek 1984). The possibility of intensive agricultural alternatives to maize production at Chunchucmil should be explored before agricultural self-sufficiency is ruled out.

The possibility of Late Classic silviculture or cultivation of other food crops, such as ramón (Brosimum alicastrum), at Chunchucmil has been discounted because of aridity and extremely thin soils. Ramón would likely have been pot-irrigated, as it is today, but preliminary field reconnaissance has revealed very few wells at the site, although they are easily dug in this environment. Furthermore, ramón would have been as precarious as maize in an unstable precipitation regime. However, stone circles at the site, about 3 meters in diameter, present intriguing possibilities. Soil may have been "quar [end p. 44] ried" from nearby wetland soils and ameliorated by mulching. These stone enclosures are often found around trees in modern gardens (Chase and Chase 1983). Based on archaeological evidence and analogy with contemporary settings in the Yucatán, crop production could have been intensified by pot irrigation and use of seed beds for crops that were later transferred to fields (Beach and Dunning, in press; Kepecs and Boucher 1991).

An additional feature at the site which may have had an agricultural function are boundary walls (albarradas), ranging in height from 20 Cm. to 1.25 m. and enclosing varying numbers of structures. The albarradas were tentatively identified as residential boundary walls by Vlcek. Low rock alignments within the albarrada groups suggest functional separation, similar to that in modern solares (kitchen gardens). Some of the walled areas are so tightly packed as to rule out large gardens. However, it is possible that these enclosed areas, like the stone circles, were used as seed beds. While our preliminary soil survey did not include testing of walled areas or areas enclosed by stone circles, we plan a systematic soil testing program, including phosphate analysis (Eidt 1984), in order to distinguish gardening from other land uses within these walls. While there are still questions about the application of the soil phosphate method (Bethell and Mate 1989), it is useful in conjunction with settlement mapping and surface ceramic surveys (Dunning 1992a, b; Killion et al. 1989).

Interesting alternative agricultural practices, such as green manuring, are being conducted in the Puuc Hills region by INIFAP (Instituto Nacional de Investigaciones Forestales y Agropecuarias), using a system variously called "slash and mulch", "slash and leave it", or tapado. Using fertilizers, a natural organic "green" manure is developed in about three years, and high maize yields have been produced (Personal communication, Darío Reyes Guerrero, INIFAP, Muna, 1994). The high humidity of the Puuc region is regarded as essential to the decomposition required for compost development (Reyes Guerrero and Aguilar Castillo 1992), and therefore the application is questionable in the Chunchucmil region. However, this level of intensification was available to the Late Classic Maya. The ability of Chunchucmil agriculturalists to adapt their techniques to widely varied environments has not been fully addressed.

ECOLOGICAL SURVEY AND SOIL TESTING
Chunchucmil occupies an extremely varied ecological and edaphic habitat. A study of the regional ecology, settlement pattern, and range of agricultural techniques requires a systematic study of this habitat, including a soil survey. Soils have been mapped at a scale of 1:250,000 by INEGI, the Instituto Nacional de Estadística, Geográfica e Informática. This level of generalization, however, is virtually useless for studies of Maya settlement. To remedy this, we conducted a two-week field survey of soils that included detailed mapping and assessment. Representative soils from each ecozone were analyzed for organic matter content, percent base saturation, and mineral content, using standard laboratory procedures of the North Central Region Agricultural Experiment Stations (North Dakota Agricultural Experiment Station 1988). The results of the soil testing program, presented below, suggest that there may be more soil variety in the Chunchucmil economic region than is immediately apparent.

The region in question contains at least six ecozones (Figure 3). The peninsular coast consists of NW-SE trending beach ridges and swales. The swamp/estuary is a 2 to 20-km.-wide zone that grades from a saline estuary on the west to a freshwater swamp on the east. The mangrove-fringed coast south of the Celestún peninsula is intersected by mangrove-lined channels ("canals") in from the coast. Several of these canals are 5 to 12 km. long, navigable by canoe, and generally oriented toward the site of Chunchucmil. Most of the canals were [end p. 45] dug in the 19th century for the extraction of logwood; some, however, may date to prehispanic times (Millet Camara 1984). The estuary grades gradually into the wet savanna (Figure 4). These two regions interdigitate over a distance of approximately 10 kilometers. Salt-tolerant species such as mangrove are gradually replaced by fresh water hydrophytic species such as cattails. The wet savanna grades eastward into drier savanna. This ecozone has been extensively cleared for shifting cultivation, henequen plantations, and livestock grazing. It includes patches of low scrub thorn forest. Dotting the estuary-savanna transition zone are many "tree islands" or petenes, formed around central cenotes that are pressently clogged with an accumulation of organic debris. They stand out from their surroundings because of the height of the canopy and the lush tropical vegetation assemblage. The petenes support dry forest species, such as dyewoods, that depend on a steady soil moisture supply (Barrera Marin 1982, Rico-Gray 1982). The cenotes fill after heavy rainfall farther inland, flooding the petenes. They have been important economically as sources of wood, chicle, palm leaves, and game. The savanna gradually grades into the karst plain, where approximately 60% of the surface is covered with limestone bedrock. The plain extends eastward to the Sierrita de Ticul, the linear fault which marks the beginning of the Puuc region. The karst plain is characterized by solution features at the surface and low (3 meters) mounds of limestone, which are called uitzes (kegelkarst hills).

CELESTUN PENINSULA
The coastal soils of the Celestún peninsula (Figure 3) were sampled at approximately 14 sites on beach ridges along a 1.25 km. transect from the coast inland. (The transect began on the coast, approximately 2 km. north of the town of Celestún). All coastal soils were calcareous, with a coarse sandy texture, and minimal horizon development. Depending on the relative age of the ridges, these soils arentisols (most recent) -or inceptisols, according to the USDA classification system, or arenosols (most recent) or cambisols, according to the FAO system (Landon 1984). Thin organic laminations occurred throughout the upper 20 cm. A horizons were predominantly dark brown sandy loam, with very low to moderate concentrations of phosphorus and low to moderate concentrations of potassium and magnesium. All had very high (70 to 97 %) base saturation of calcium and, usually, low organic matter content in the A horizon (1 to 11 %). Parent material was coarse calcium carbonate sand and shells.

An inland chronosequence of deeper and more developed A horizons and generally greater pedogenesis is found on the successively older, stabilized beach berms. The beach ridges farthest inland had high percentages of organic matter, elevated levels of phosphorus and magnesium, and significantly higher cation exchange capacities. The texture of these soils is noticeably siltier. The fertility of the inland ridges is evident in the modern milpa cultivation there (Figure 5). Two milpas, one recently burned and one fallow, were sampled. The milpa soils had dark brown sandy loam A horizons, with small granular structure, and many roots. The AB horizon was a lighter brown color, a loamy sand, still having many roots and some shell fragments. The parent material was coarse calcium carbonate sand, and a caliche was evident at 94 cm. This soil had a high per cent base saturation with calcium, and rather low concentrations of phosphorus, potassium, and magnesium. (The fallow field had a markedly lower mag [end p. 46] nesium concentration than the recently prepared milpa, presumably due to uptake by crops.) We classsified these inland ridge soils as haplustolls (USDA) or haplic kastanozems (FAO).

ESTUARY AND SAVANNA
The petén that we sampled (Figure 3) had very dense forest cover and a thick understory. Depth to the water table in the petén was 44 centimeters, during an unusually dry wet season in 1994. We classified this soil as a fibric troposaprist (USDA) or histosol (FAO). The top 15 centimeters consisted of an Oe or Oj horizon with an organic matter conntent of 41%, and relatively higher magnesium and lower calcium concentrations than the surrounding savanna soils. Below the surface horizon was a peaty 0 horizon with occasional limnic horizons. The extent to which the Maya used these areas is unknown.

Local farmers refer to the soil in the petenes as boxluum, which is a deep, dark brown, highly organic soil found in the bottom of sinkholes (Dunnning 1992a). Local informants reported that the milperos who cultivate the petenes live there for the season to guard their crops from animals and intruders because the predominant cash crop grown in the petenes is marijuana.

In the wet portion of the savanna, it is difficult to find soil deeper than 5 cm. In one sample, total soil depth was 20 cm., consisting of an O horizon overlying a limnic or diatomaceous layer. The surface was covered by a thin white crust. Although there is a great deal of variability in the savanna soils, a typical soil profile has an approximately 10 cm. deep A horizon, subangular blocky structure, and a B horizon with clay skins on the peds. About 50% of the surface is covered with bedrock and cobbles. Hydrophytic vegetation occupies low areas, and grasses occupy slightly raised areas. The wooded portion of the savanna is composed of a low thorn forest of acacia, cacti, and palms.

We sampled one small bajo (a small, gently slopping depression) in the savanna. The soil was a histosol (USDA, FAO) with a deep black 0e horizon, and abundant leaf litter and hemic material. The A horizon was thin, dark, and granular, overlying a grey moist clayey layer, mottled with white patches.

The savanna region is used for small milpas and pastures. Maize is grown amidst cobbles of limestone heaped in small mounds less than one meter high, but elevated from inundation in the wet season (Figure 6). The flat area between mounds is not used for maize cultivation due to grass invasion, according to a local informant. The soil in these low areas is up to 20 centimeters deep, and has a reddish brown A horizon, with a subangular blocky struccture. The milpero informant referred to this soil as kancab, which generally refers to a low-lying, heavy, red-yellow soil, an ustalf (USDA) or calcic luvisol (FAO; Dunning 1992a). In solutional cracks in the limestone, there was a soil 21 cm. deep in places, dark reddish brown, with predominantly clay texture and clay skins on the peds in the lower portion of the C horizon, of which this soil is entirely composed. A local informant told us that he probes the ground with a metal rod to find these deep spots for planting maize. The soil in the cropped area was well [end p. 47] -aerated, had a subangular blocky structure, and conntained many rock fragments and charcoal. It had elevated levels of phosphorus and magnesium, possiibly due to fertilization (Figure 7).

KARSTIC PLAIN
Soils and geomorphology change gradually as the savanna grades into the karst plain. We estimate that limestone covers about 60% of the surface, and uitzes increase to several meters in height. This area, where the core of Chunchucmil is located, is covvered by abandoned henequen plantations and pasture. Soils are thin, orange to reddish brown, with granular structure and many fine roots and limestone nodules throughout. The soils consist of an A horizon, approximately 10 to 15 centimeters deep. Exceptions are pockets of deeper soils, found in solutional cracks in the limestone, where soil is up to 30 cm. deep and shows a surface horizon of subangular blocky structure and clay skins in macropores at depth. The soils in the cracks show no effervescence throughout the profile, probably because of the high degree of leaching. Many of the uitzes in this area are being cultivated for maize. The low spots between mounds often contain rich, relatively deep mollisols, which can be extremely ferrtile soils with proper management. A typical mollisol in this area has a total depth of over 30 cm., a dark brown upper Al horizon with fine granular structure, clay loam texture, an organic matter content of 11 %, many roots, and abundant earthworms.

In the thorn forest, which covers the uncleared portion of this region, we sampled a soil that had an OJ horizon of partially decomposed organic matter, a reddish brown A horizon, and a B horizon. The infiltration rate was slightly higher in the forested soil than in the open pasture. In one abandoned henequen field in this zone, the infiltration rate was as low as 5.3 mm. per minute. Presumably, the soils in the cleared open pasture portion of the site have been greatly altered and compacted. The A horizons had organic matter contents of about 10.5% and potassium, calcium, and magnesium concentrations similar to the other savanna soils. We classified these soils as lithic rendolls (USDA) or rendzinas (FAO).

In summary, our preliminary soil survey concludes that the coastal zone is generally not suited for cultivation with the exception of inland beach ridges, which are limited in areal extent. In the estuary and estuary-savanna transition zone, the petenes contain histosols, which are potential sources of organically-enriched soils for mulching. The agriicultural potential of the seasonally inundated savanna is unclear; the lower portions of the savanna are inundated during the rainy season and support largely hydrophytic vegetation. Kancab is a fairly fertile soil if properly managed (Dunning 1992a), but grass invasion limits its agricultural usefulness for modem milperos. The histosols in the bajos in the savanna are additional potential sources of organic soils for mulching. The soils of the karstic plain are of highly varied potential, ranging from very thin soils to deep, rich, fertile mollisols.

CONCLUSIONS
"Chunchucmil" is a Yucatec Maya term, meaning "beneath the branches of the Chukum tree". A brief look at the soil beneath the roots of the Chukum tree has revealed that the common description, "generally thin soils", does not adequately portray the full range of soil types and potentials in the economic zone of Chunchucmil. In addition, the assessment of the past agricultural potential of this region is clouded by the historical land uses of pasturing, henequen cultivation, and orchard cropping, all of which can be destructive land uses. This is evident in the severe soil compaction in modern pastures. The potential for ancient gardening and other means of intensification, such as mulching and seed bed cultivation, is worth further investigation. A program of mapping and soil analysis at the site to determine the function of structures such as stone circles and rock alignments is necessary before this solution to "the Chunchucmil problem" is ruled out.

Preliminary indications are that, while the agricultural potential of the site and its environs may have been more promising than is indicated by present day landuse and productivity, the ability of the site to sustain the population density remains unlikely. Whether salt was an important trade commmodity, and whether nearby regions, such as the Puuc, had enough surplus to supplement the food supply of Chunchucmil are, as yet, unanswered questions. An extensive program of mapping, surface survey, and soil testing is needed to fully answer the Chunchucmil question. [end p. 48]

Acknowlegements
We would like to thank the National Science Foundation and Howard University for funding the soil survey during the 1994 field season.

REFERENCES
Andrews, Anthony P. 1983. Ancient Salt Production and Trade. Tucson: University of Arizona Press.

__________. 1990. "The role of trading ports in Maya civilization". In Vision and Revision in Maya Studies, Flora Clancy and Peter D. Harrison (eds), pp. 159-168. Albuquerque: University of New Mexico Press.

Barrera Marin, Alfredo. 1982." Los petenes del norte de Yucatán: su exploración ecológica en perspectiva", Biotica 7(2): 163-166.

Beach, Timothy and N. P. Dunning. 1995. "Ancient Maya terracing and modern conservation in the Peten rain forest of Guatemala".i> Journal of Soil and Water Conservation 50:138-145.

__________ . In press. "An ancient Maya reservoir and darn at Tamarindito, El Peten, Guatemala", accepted by Latin American Antiquity.

Bethell, P. and I. Mate. 1989. "The use of soil phosphate analysis in archaeology: a critique". In Scientific Analysis in Archaeology and its Interpretation, J. Henderson (ed), pp . 1-29. Monograph #19, Oxford University Committee for Archaeology. Oxford, U.K.:Archaeological Research Tools #5.

Carta de Evapotranspiración y Deficit de Agua, 1:1,000,000, Mérida. 1983. INEGI (Instituto Nacional de Estadística, Geográfica e Informática).

Carta de Precipitación Total Anual, 1: 1, 000, 000, Mérida. 1981. INEGI (Instituto Nacional de Estadística, Geográfica, e Informática).

Chase, Arlen F. and Diane Z. Chase. 1983. "Intensive gardening among the Late Classic Maya". Expedition, Spring, 1983, pp. 2-11.

Dahlin, B. H., M. Brenner, B. Leyden, and J. Curtis. 1991. "An 8000-year Record of Paleoclimate from the Northwest Yucatan, Mexico". Paper delivered at the Conference on Ancient Maya Agriculture and Biological Resource Management, University of California, Riverside, August 22-24.

__________ .1992. "Preclassic and Early Classic Period Climate Change at San José Chulchaca, N.W. Yucatan, Mexico". Paper delivered at the Annual Meeting of the Society for American Archaeology, Pittsburgh.

Dunning, N. P. 1992a. Lords of the Hills: Ancient Maya Settleement in the Puuc Region, Yucatan, Mexico. Monographs in World Archaeology, #15. Madison: Prehistory Press.

__________ . 1992b. "Umwalt, Siedlungsweise, Ernahrung und Lebensunterhalt im Maya -Tiefland wlihrend der Klasik (250-900 n. Chr.)." In Die Welt der Maya. Mainz am Rhein: Verlag Philipp Von Zabern.

Eidt, R.C. 1984. Advances in Abandoned Settlement Analysis: Applications to Prehistoric Anthrosols in Colombia, South America. Center for Latin American Studies. Milwaukee: University of Wisconsin.

Flannery, K. (ed). 1982. Maya Subsistence: Studies in Memory of Dennis E. Puleston. New York: Academic Press.

Gómez-Pompa, S. Flores, and V. Sosa. 1987." The Pet-Kot: A man-made tropical forest of the Maya". Interciencia 12:10015.

Harrison, P. and B. L. Turner, II (eds). 1978. Pre-Hispanic Maya Agriculture. Albuquerque: University of New Mexico Press.

Hodell, David, Jason Curtis, and Mark Brener. 1995. "Possible role of climate in the collapse of the Classic Maya civilization." Nature 375: 391 - 394.

Kepecs, Susan, and Sylviane Boucher. 1991. "The Prehispanic Cultivation of Rejolladas and Stonelands: New Evidence from Northwest Yucatan".ln The Managed Mosaic: Ancient Maya Agriculture and Natural Resource Use, Scott Fedick (ed). Salt Lake City: University of Utah Press.

Killion, Thomas, Jeremy Sabloff, Gair Tourtellot, and N. P. Dunning. 1989. "Surface assemblages at Terminal Classic (A.D. 800 - 1000) Sayil, Puuc Region, Yucatan, Mexico." Journal of Field Archaeology 16: 273-294.

Kurjack, E. B. and S. Garza 1981. "Precolumbian Community Form and Distribution in the Northern Maya Area." In Lowland Maya Settlement Patterns, W. Ashmore (ed), pp. 287-310. Albuquerque: University of New Mexico.

Landon, J. R. (ed). 1984. Booker Tropical Soil Manual. Bath: Pitman Press Ltd.

Leyden, B., M. Brenner, J. Curtis, D. Piperno, T. Whitmore, and B. Dahlin, in press. "A record of long and short-term cliimatic variation from Northwest Yucatan: Cenote San José Chulchaca." In Proceedings of the Conference on Ancient Maya Agriculture and Resource Management, University of California, Riverside. Salt Lake City: University of Utah Press.

Millet Camara, Luis. 1984. "Logwood and archaeology in Campeche." Journal of Anthropological Research 40(2): 324-328.

North Dakota Agricultural Experiment Station. 1988. Recommmended Chemical Soil Test Procedures. North Central Regional Publication No. 221 (Revised). Fargo: North Dakota State University.

Pohl, M. D. (ed). 1990. Ancient Maya Wetland Agriculture. Boulder: Westview Press.

Reyes Guerrero, Darío, INFAP, Muna, 1994. Personal Commmunication.

Reyes Guerrero, Darío, and Guillermo Aguilar Castillo. 1992. "lntensificación de la milpa en Yucatan". In La Modernización de la Milpa en Yucatan: Utopia o Realidad. Merida: Centro de Investigación Científica de Yucatan.

Rice, Don S. 1993. "Eighth Century Physical Geography, Environment, and Natural Resources in Maya Lowlands." In Lowland Maya Civilization in the Eighth Century AD , J. A. Sabloff and J.A. Henderson (eds), pp. 11-63. Washington, D.C.: Dumbarton Oaks.

Rico-Gray, Victor. 1982. "Estudio de la vegetación de la zona costera inundable del noroeste del Estado de Campeche, México: Los Petenes". Biotica 7(2): 171-190. [end p. 49]

Roys, R. L. 1957. The Political Geography of the Yucatan Maya. Carnegie Institute of Washington, D.C. Publication # 613.

Shuman, M. 1974. The Town where Luck Fell. Ph.D. dissertation, Department of Anthropology, Tulane University.

Turner, B. L. and C. H. Miksicek. 1984. "Economic plant species associated with prehistoric agriculture in the Maya lowlands." Economic Botany 38:179-193.

Vlcek, David T., Sylvia Garza de González, and Edward B. Kurjack. 1978. "Contemporary farming and ancient Maya settlements: some disconcerting evidence." In Prehispanic Maya Agriculture, P. D. Harrison and B. L. Turner II (eds), pp. 211-223. Albuquerque: University of New Mexico Press.

Wellhausen, E. J. 1957. Races of Maize in Central America. National Academy of Sciences, National Research Publication # 511.

Whitmore, T. J., M.Brenner, J. H. Curtis, B. W. Leyden and B. H. Dahlin. 1994. "Holocene Climatic and Human Influences on Lakes of the Yucatan Peninsula, Mexico." Paper presented to the International Limnological Society, Milan, August, 1994.

Wilson, E. M. 1980." Physical Geography of the Yucatan Penninsula." In Yucatan: A World Apart, E. Mosley and E. Terry (eds), pp. 5-40. Birmingham: University of Alabama.

RESUMEN
Ubicado en el noroeste de la Peninsula del Yucatán, Chunchucmil fue el centro Maya con la mayor densidad poblacional durante el Período Tardío Clásico. La alta densidad poblacional de este sitio, continua como un misterio, ya que el potencial agrícola de dicho sitio es bajo. Actualmente este sitio es extremadamente seco y la superficie está cubierta por una extensa capa de piedra caliza mezclada con una capa firma y escasa de suelo. Las siguientes son posibles respuestas a la paradoja de Chunchucmil: I) Los agricultores de Chunchucmil utilizaron técnicas de intensificación agrícola para aumentar la producción y 2) Chunchucmil estaba envuelta en un comercio regional de recursos costaneros de comida (tales como el sal). Entender el patrón de colonización, fronteras del sitio, y la base substancial de Chunchucmil requiere un extensivo programa de colonización y delineación, en adición a una medición sistemática del suelo. A través de un programa extensivo de investigación, se condujo una medición ambiental preliminar del suelo de esta región en 1994. Este trabajo presenta los resultados de dicho análisis, alternativas para prácticas agrícolas, y sugerencias para futuras investigaciones del sitio. [end p. 50]