Mapping La Milpa: a Maya city in northwestern Belize.

Tourtellot, Gair, III ; Clarke , Amanda ; Hammond, Norman 等

Introduction

La Milpa is a major Lowland Maya site in northwestern Belize, close to the frontiers with the Guatemalan Department of El Peten and the Mexican State of Quintana Roo, and in an area until recently relatively inaccessible because of the lack of roads through the rainforest. The research reported here, complementing that by several other projects in the region, follows the acquisition of some 350,000 acres of forest by the Programme for Belize as an ecological preserve and the establishment of a Scientific Advisory Board for Archaeology.

The Boston University-National Geographic Society La Milpa Archaeological Project (LaMAP) began field operations in February 1992: our aim is to build up an holistic picture of an ancient Maya community that appears to have flourished for several centuries, by mapping the pattern of settlement and its relationship to landscape, studying the range of mineral and plant resources available to the inhabitants and by carrying out surface collection and excavations to determine the nature, extent and persistence of Maya culture there and understand the ancient city in its environmental context.

Because of the location of La Milpa in a biosphere reserve, our methods are designed to have the minimum impact on the present forest environment: large-scale stripping of vegetation and topsoil from the temples and plazas of the site core, followed by consolidation or restoration of numerous buildings, such as has been carried out on some other Maya sites in Mexico, Guatemala and Belize, is not part of our remit. Such exposure of buildings as we do carry out over the next decade must reconcile the legitimate demands of archaeo- and eco-tourism with the long-term conservation of La Milpa.

The site of La Milpa

La Milpa is located on the eastern edge of the core area of Classic Maya civilization, centred in the northeastern Peten and adjacent areas of Mexico. It lies mid-way between the well-known and recently investigated sites of Rio Azul (Adams 1990) and Lamanai (Pendergast 1981), both with a long history of occupation, from perhaps as early as the late Middle Preclassic (600-400 BC) onwards. The city occupied an upland area between the Rio Bravo escarpment to the east and south, and the Rio Azul to the north; the western side of the site is drained by an intermittent tributary of the Rio Azul which we have named Thompson's Creek. The centre of La Milpa stands about 180 m above sea level (asl) on a prominent limestone ridge, the highest point for some distance around. Rugged forested terrain stretches west into Guatemala and east to the Rio Bravo, where the scarp drops from 160 m asl to the coastal plain at 20 m asl within 2 km.

Loran coordinates for La Milpa are given by Ford & Fedick (1988: 15) as 17 |degrees~ 49 16N and 89 |degrees~ 0321W. We independently obtained a location approximately 1 km farther north at 17 |degrees~ 5006N, 89 |degrees~ 0306 (UTM 16Q BQ 2-82-637E, 19-72-929N), using a hand-held Magellan GPS receiver atop Pyramid 1 on the Great Plaza. The research area of LaMAP is defined as a 6-km radius from the Great Plaza: comparison with existing maps of Tikal, Seibal and other sites suggests that the entire community, including several subordinate ceremonial precincts, should lie within this 113 sq. km.

Previous work

La Milpa was first explored and named by the late Sir Eric Thompson in March 1938. He mapped the main plaza and the locations of 12 stelae therein, and noted the complex of courtyards south of the plaza (Thompson 1938; Hammond 1991). Stela 7 bore a clear CR date of 12 Ahau 8 Pax and the corresponding IS of 9.17.10.0.0. (30 November AD 780), and the other carved stelae Thompson was able to observe were also Late Classic in style. Thompson thought that the site might be the largest in British Honduras (as Belize then was; neither Caracol nor Lamanai had been explored at this point), but the only other published reference to it (Thompson 1939: 280) simply lists the 'pyramids, mounds, enclosed courts, nine sculptured and three plain stelae, plain altars'.

No further exploration took place until David M. Pendergast and H. Stanley Loten paid a brief visit in the 1970s, and made a pace-and-compass map of the main plaza (unpublished); M. Gutchen and E. Luna of the Department of Archaeology in Belmopan visited La Milpa in 1979 to follow up reports of looting, which were unfortunately true; further Department of Archaeology visits took place in 1985, the first with the Police Tactical Services Unit to check on reports of extensive looting and marijuana fields (both true): by then all the main structures around the Great Plaza and numerous smaller ones had been penetrated by large trenches. Some of these revealed earlier buildings, and some had struck the tombs they were seeking, but many had done little beyond destabilising the rubble fill of the structures.

The first archaeological exploration since Thompson's time was carried out in 1988 by two separate groups: the Rio Bravo Project made a sketch map of part of the site centre (Guderjan 1989), and the Programme for Belize commissioned an evaluation of the archaeological resource potential and management of the lands it was acquiring in northwestern Belize (Ford & Fedick 1988). Ford & Fedick mapped the main plaza area and the enclosed courtyards to the south, including the locations of stelae and looters' trenches. Guderjan (1989: figure 5) combined this with his own 1988 data to produce a composite map of the main plaza that corrected the orientation of some structures and provided numerical identifiers for the structures (I-XVI) and stelae. Ford & Fedick (1988: figure 6) also mapped a 1-ha sample of settlement north of the centre, finding a density equal to 240 strs/sq. km.

In February-April 1990 Guderjan carried out extensive further mapping, including four plazas, 20 residential courtyards and 85 structures in a 1 km X 0.5 km area of the site core. Looters' trenches were investigated and profiled where appropriate (Guderjan 1991: figures 3-8). Guderjan estimates that La Milpa would have at least 24 and perhaps up to 30 courtyards, which on Adams & Jones' (1981) scale would place the site in the topmost rank.

A visit by Hammond in March 1990 resulted in the identification of one fragmentary monument (Stela 1) as Early Classic in date, and in corroboration of the impression obtained from Thompson's field notes that much of the major architecture was likely to be Preclassic in origin; a visit by Vernon Scarborough (University of Cincinnati) resulted in the identification of two large depressions south of the main plaza as reservoirs (although they probably originated as the quarries from which the construction fill for the plaza was dug), as well as other water management features; these were investigated in 1992 and will be reported on separately by Scarborough.

Overall, the work carried out at La Milpa through 1990 resulted in the clearing and mapping of the bulk of the site core as initially explored by Thompson, and in brief descriptions of its structures and monuments (Guderjan 1991: 7-34); Guderjan's numbering of the stelae in the Great Plaza (Plaza A) unfortunately differs from that established by Thompson (1938) and cited by others (e.g. Proskouriakoff 1950: figure 64c). LaMAP uses Thompson's numbering of Stelae 1-12, with the subsequently discovered monuments as Stelae 13-16.

Investigations in 1992

The mapping and excavation programmes pursued by LaMAP in 1992 were an attempt to understand the structure and chronology of the ancient community. The three basic goals of the mapping, directed by Tourtellot, were

1 to acquire an understanding of social organization based on the variety, typology and distribution of architectural assemblages;

2 to match this with information on economic and political organization, including sectorial development, location of production facilities and communications such as causeways, and definition of community boundaries and satellite centres;

3 to understand how the rugged and diverse environment of the tropical rainforest was exploited in terms of agriculture, silviculture and the associated management of land and water resources.

Procedures can conveniently be divided into survey and mapping: the former used an electronic distance measuring instrument (EDMI), supplemented with Lietz optical transits, to align and stake out a grid of trails (brechas) cut through the undergrowth at 500-m intervals. These were highly controlled for azimuth, distance and altitude, and the initial set of brechas formed a 1 X 1 km cruciform within a 1 X 1 km square, giving four closed cells each 500 m on a side; most of this central 1 sq. km was mapped in detail in 1992. The arms of the cross will be extended outwards in 1993 and lateral brechas run from them. The baseline was laid out on 1992 magnetic north along the east side of the Great Plaza, just west of the line of stelae that stand in front of the major pyramids; it was extended south between Structures 3 and 8, and the main east-west cross brecha placed where it would not have to pass over tall buildings.

This intersection was designated E6000/N6000 in the survey grid, which had a hypothetical southwestern origin beyond the 6-km radius of the project's research area, and became the Point of Beginning (POB) for all subsequent work. An altitude of 180 m asl was assigned to the POB on the basis that the 180-m contour on the Belize 1:50,000 map (D.O.S. 4499, Sheet 8, 1976) enclosed the area of the site core, and that this was somewhat corroborated by several GPS readings of just over 190 m asl for the top of Structure 1 on the Great Plaza (which lies at 215.8 m asl on the basis of the datum used here).

Stakes (in most cases recycled aluminium tent poles cut into 0.5 m sections) were sunk into the ground at 100-m intervals along the main brechas to serve as tie-ins for more detailed mapping. In addition the Great Plaza was staked on a 25-m grid to check the size and orientation of structures on the earlier maps made by Thompson (Hammond 1991: figure 1) and Guderjan and Lindeman (Guderjan 1991: figure 2), and provide convenient datum points for excavations. Altitudes on stakes and intermediate and lateral ground points were used to contour the natural topography.

Mapping, as distinct from survey, used the compass and pace technique developed at Tikal (Puleston 1983: 4-9). This uses teams systematically walking back and forth in parallel along narrow transects on compass headings, with the team leader using calibrated pace, traverse compass and hand level to measure all cultural and natural features reported by team members. In 1992 teams consisted of a surveyor, flanking student observers and a Belizean worker clearing the way with a machete; each team sweep covered a transect about 75 m wide. We used this method because of its speed: no sight lines needed to be cleared for an EDMI.

Transect bearings and building orientations were determined with Suunto KB14/360 |degrees~ traverse compasses, elevations from a hand level mounted on a 1.5-m staff. A pace was calculated at 0.7 m, and the length of each transect was retroactively calibrated by dividing the 500 m between brechas by the number of paces taken. Most structures were plotted in the field, allowing the map to be checked for ground truth. Accuracy in detail is fair, especially at the 1:1000 scale of plotting where the thickness of a pencil line can equal 0.5 m on the ground and given the imprecise edges of most ruined Maya structures under bush cover; ground plans as plotted tend to be more regular than the actual lines of collapse. Wall stones at La Milpa tend to be heavily eroded and disordered, and structural lines were usually established from a combination of contour breaks and toppled walls or exposed fill. More precise means than compass and pacing are needed where exact orientations, dimensions and volumes are needed, but the methods we used did not run the risk of being more precise than the data available warranted. In either case, accuracy depends more on the archaeologist than the instruments used.

Gross control was provided by the staked brechas into which the mapping transects tied, so that the maximum possible locational error under dense bush was |approximately~ 15 m. Such mapping is highly efficient for finding and identifying ruins, plotting their gross orientation and spacing, numbers and clustering, approximate sizes and floor plans and association with topography. Natural features mapped included major changes in contour, elevations such as rock outcrops and rock shelters under overhangs, depressions which may be solution hollows and streambeds. Cultural features apart from the mounds of collapsed buildings included spreads of rocks (perhaps construction stockpiles), possible quarries and underground chultun chambers cut into bedrock.

Site layout

The ceremonial precinct of La Milpa lies on a ridge about a kilometre across and more than that in length from north to south; its summit at c. 195 m asl is just north of the Great Plaza, and the ridge slopes gently southwards and more steeply to east and west. The eastern side, with steep drops of up to 20 m, is heavily dissected by the channels of seasonal streams, bordered by exposures of limestone caprock over softer marl which has often undercut to form rockshelters. Late/Terminal Classic bowl and jar fragments from these suggest their use for the collection of zuhuy ha, ritually-important 'virgin water', issuing magically from inside the rock. A quarry for high-quality limestone existed along one of these small gorges, and stone of this type was used for the carved stelae erected in the Great Plaza (James R. Ashby pers. comm.. An aguada (waterhole) 100 m across was found beyond the southwest end of the ridge, holding water well into the dry season; this is the only permanent water source so far known at La Milpa, although two reservoirs (initially quarries) lie just south of the Great Plaza and were investigated by Vernon Scarborough in 1992. The satellite centre of Say Ka is some 4 km to the southwest, within the posited settlement zone of La Milpa; other sites lie on the ridges beyond, and are being investigated by a project from the University of Texas directed by R.E.W. Adams and Fred Valdez, Jr.

The public architecture of La Milpa occupies an area some 680 X 250 m along the ridge, with two principal areas of construction. In the north the Great Plaza, at c.165 X 120 m (about 20,000 sq. m) one of the largest public spaces ever built by the Maya, is dominated by four large temple-pyramids (Strs. 1-3, 10). South of this are the two reservoirs, and beyond them the second main group of buildings, consisting of three open plazas and a series of enclosed courtyards, with only one pyramid and no stelae. No excavations were carried out in this southern area in 1992: our efforts were concentrated in the northern sector. The map of the Great Plaza area used here shows the conventions used in Maya archaeology: mounds are depicted as geometric solids with a diagonal and inner rectangle proportional to the height and top surface area. In many cases the top surface would have supported a walled superstructure, but only where (as on Str. 8) the outlines of the rooms are fairly clear even below vegetation and topsoil is this shown. Whether the retaining walls of substructures were battered (as the convention implies), terraced or vertical can only be established by excavation: all three types are exposed in the sections of looters' trenches at La Milpa.

The plaza is dominated by the line of three pyramids along its eastern side, and by Str. 10 within the plaza itself. Str. 1, the tallest, stands 24 m above the plaza floor; its flat top and rough masonry suggest a Late Preclassic date similar to structures at El Mirador and Nakbe in the Peten (Graham 1967). Several looters' tunnels reveal earlier buildings within Str. 1 (Guderjan 1991: figure 3), and later ones were added at plaza level against its west front. None of the other pyramids has been investigated as yet, although the location of Str. 10 on the same axis as Str. 1 suggests that they may be coeval. The small Str. 5 in front of the gap between Strs. 1 and 2 should be later, although it also has at least two phases of construction visible in looters' trenches; one of these may be coeval with Stela 7 of AD 780, which stands directly in front of Str. 5. The South Ballcourt, Strs. 6 and 7, similarly blocks part of the view of Str. 3 from the plaza, suggesting that it is a later insert, while the North Ballcourt, Strs. 11 and 12, lies athwart the main entry into the Great Plaza from the north and is known from excavation to be of Terminal Classic date.

The south side of the plaza is occupied by a single range structure 90 m long, Str. 8, with a collapsed masonry superstructure, apparently of 13 rooms. The number was significant to the Classic Maya, there being 13 oxlahuntiku, gods of the heavens, who stood in opposition to the 9 bolontiku, lords of the night and the underworld. Adjoining it on the west is a raised 'acropolis' with an enclosed courtyard some 6 m higher than the plaza and cut off from it by the bulk of Str. 9 rising to a height of 14 m. The buildings on the courtyard (Strs. 13-15) are more modest in scale and probably formed a royal residence compound, defensible and difficult of access. The northwestern and northern sides of the Great Plaza are bordered by low substructures, a curious contrast to the massive architecture of the other sides; the location of the large Str. 3, at 75 X 50 m in basal area the bulkiest of the pyramids, half off the plaza facing the end of Str. 8 is also odd.

Monuments

Most of the 16 stelae so far found at La Milpa line the eastern side of the plaza in front of the pyramids. Thompson (1938) recorded 12, most of them still in situ; Stela 4 is probably buried under looters' rubble just north of Stela 5, and Stelae 9 and 10 have not yet been disentangled from the pile of fragments at the northwest corner of Str. 3 (Guderjan (1991: 33-34) here has one stela, his #13, and an altar). Stela 1 was turned by one of us (NH) in 1990 and identified as being of Early Classic date (AD 250-600) on stylistic criteria. Of the others, Stelae 2 and 5, although broken and eroded, may be Late Classic (AD 600-800) in date, as are certainly the better-preserved Stela 8 (still standing in front of Str. 2) and Stela 12 (recumbent in front of Str. 3). Stela 7, the only one with a well-preserved inscription, was dated by Thompson (1938) to 9.17.10.0.0. 12 Ahau 8 Pax (30 November AD 780 (11.16. correlation at 584,283)). The Initial Series date on the north edge is confirmed by repetition of the Calendar Round date on the west front; both the south edge and the cruciform glyph panel on the back should be readable with proper lighting, as may be the two glyphs on the front at lower right which seem to name the ruler portrayed.

Stela 12 has survived erosion better on its lower part and side, where the legs and elaborate ex loincloth with its squared spirals and the left arm outstretched holding a circular shield are clear, than on the upper, where the profile face is carved as inset relief and the right hand holds an upright spear decorated with stiff lozenges. Some of the glyphs on the north edge may be readable.

Two other stelae lie in the Great Plaza: Stela 13 stood opposite the east end of the North Ballcourt, and although tall (3.45 m) was uncarved. The tiny Stela 16, lying loose, beyond the west end of the ballcourt beside a much larger and fragmented altar, was carved in Early Classic style with a standing figure rather similar to that on Stela 1. Both of these monuments were noted by Guderjan (1991: his #8 and 10), as were the two monuments outside the Great Plaza, Stelae 14 and 15 (his #11 and 12). Of these, Stela 14, standing in front of the west side of Str. 16 but actually facing northwards towards Str. 3, is both plain and broken into three; the butt and the altar in front are still in situ.

Stela 15, 3.42 m high, was found lying face down in front of a small and heavily-looted mound south of Reservoir A, adjacent to two linear structures which may frame a southern approach to the Great Plaza. Glyph panels on the edges have been eroded, but when the stela was raised to a vertical position, night photography revealed a striking figure on the west face. Standing in right profile, with well-defined chest, back, buttocks and thighs, he wears what seems to be a small round shield on his shoulder with one or more spears tucked behind it. His head and face are disproportionately large, with a prominent nose; part of the elaborate headdress hangs down in front of his face. A small panel of incised glyphs in front of the shins may be readable, but stylistically Stela 15 looks to be late in Maya history.

Overall the stelae of La Milpa, of which we found eight to be carved (Thompson thought another four were, now too eroded to tell) and three plain, span a period of at least two centuries, c. AD 580-780, and possibly double that. While many of the late monuments were more or less complete (Stelae 5, 7, 8, 12, 15) and in some cases still standing (Stelae 7, 8), the earlier Stela 1 had been broken off midway and the upper half removed elsewhere. Stelae 2 and 6 were similarly mutilated: all three monuments stood in front of Str. 1 (although Stelae 3 and 5 there seem to be undamaged). What political events in La Milpa's history this reflects remains to be ascertained.

That this history was a long one is clear from the results of our initial excavations and surface collections. A test pit in the centre of the Great Plaza (Op. A01) showed that deep Late Preclassic construction deposits underlay the final levelling-up of the plaza surface in Late Classic times; another (Op. A03) north of the elite acropolis compound west of the plaza, yielded almost entirely Late Classic and Tecep (Terminal Classic-Early Postclassic: AD 800-1100) material, as did superficial excavations around the North Ballcourt. Surface collections, from the settlement as well as in the site core, were predominantly of Tecep date, constituted of heavy storage jars and basins of a coarse calcite-tempered paste. The same range of vessels was found in the excavation of two residential courtyards just northeast of the Great Plaza (Ops. A06 and A07; Guderjan 1991: figure 2, #48, 49). Their typology links the late ceramics of La Milpa to Peten sites further west and south, rather than to Lamanai and Nohmul to the east and north, with their strong links to the Yucatan Peninsula: La Milpa seems to be at the northeastern limit of Peten regional culture, something confirmed by the site's architecture and sculpture.

Some northern influence is apparent at the very end of La Milpa's history, however: a rectangular building with metre-wide dwarf walls was constructed in the southwestern corner of the Great Plaza, arguably when it had ceased to be a functioning ceremonial space, and investigated as Op. A04. The abundance of domestic pottery indicated that this was a residence, and its closest architectural parallels are Strs. 132, 139 (Hammond 1985: figures 4.6-4.29), and 141 in the northern suburbs of Nohmul; these in turn have strong architectural and ceramic links with Tecep and Postclassic Yucatan. Given the northern influence, perhaps even invasion, seen at this time at Rio Azul by Adams (1990: 35), where there is also a late monument (Stela 4) possibly comparable with La Milpa Stela 15, it seems likely that we are here close to the southernmost limit of that movement of Yucatecan culture, be it by invasion, migration or simply emulation.

Acknowledgements. Funding for LaMAP was provided by the National Geographic Society, by Boston University, and by donations from Francis Ford Coppola and an anonymous benefactor. Permission to work at La Milpa was granted by the Government of Belize through the Department of Archaeology, where we are most grateful for John Morris and his staff. Programme for Belize offered the use of research station facilities and the fullest cooperation: we are most grateful to Arnold Brown, Joy Grant, Peter Herrera and John Masson. Kathy Tips and Francisco Estrada Belli supervised excavations; these and mapping were carried out by students from Boston University and the University of Texas at San Antonio. Laboratory work was directed by Jim Spriggs and Sarah Kingsley, and drafting by Sheena Howarth and Sarah Kingsley; logistics were managed by Mark Hodges and Ovel Martinez, and the kitchen by Argelia Martinez. Our colleagues Richard E.W. Adams, Thomas H. Guderjan, Vernon Scarborough, and Fred Valdez, Jr, who are also working in and around La Milpa, have been hospitable collaborators. We owe diverse debts of gratitude to Neva Folk, Mary Ann Harrell, Yvonne Keast, Evelyn Y. LaBree, Mary Griswold Smith and George E. Stuart.

References

ADAMS, R.E.W. 1990. Archaeological research at the lowland Maya city of Rio Azul, Latin American Antiquity 1: 23-41.

ADAMS, R.E.W. & R.C. JONES. 1981. Spatial patterns and regional growth among Classic Maya cities, American Antiquity 46: 301-22.

FORD, A. & S. FEDICK, 1988. Draft report on the archaeological resource potential and management of the Programme for Belize Lands, Orange Walk, Belize. Santa Barbara (CA): MesoAmerican Research Center, Social Process Research Institute, University of California.

GRAHAM, I. 1967. Archaeological explorations in El Peten, Guatemala. New Orleans (LA): Middle American Research Institute, Tulane University. Publication 33.

GUDERJAN, T.H. 1989. An archaeological reconnaissance in Northwestern Belize, Mexicon 11: 65-8.

1991. (Ed.). Maya settlement in Northwestern Belize: the 1988 and 1990 seasons of the Rio Bravo Archaeological Project. Culver City (CA): Labyrinthos.

HAMMOND, N. (ed.). 1985. Nohmul: a prehistoric Maya community in Belize. Excavations 1973-1983. Oxford: British Archaeological Reports. International series 250.

1991. The discovery of La Milpa, Mexicon 13: 46-50.

PROSKOURIAKOFF, T. 1950. A study of Classic Maya sculpture. Washington (DC): Carnegie Institution. Publication 593.

PULESTON, D.E. 1983. The settlement survey of Tikal. Philadelphia (PA): University Pennsylvania Museum. Tikal Report 13, University Museum Monograph 48.

THOMPSON, J.E.S. 1938. Reconnaissance and excavation in British Honduras. Pp. 16-17 in Annual Report of the Division of Historical Research, 1937-38, Carnegie Institution of Washington Year Book 37: 1-37. Washington (DC): Carnegie Institution.

1939: Excavations at San Jose, British Honduras. Washington (DC): Carnegie Institution. Publication 506.