No slow dusk: Maya urban development and decline in La Milpa, Belize.

Hammond, Norman ; Tourtellot, Gair ; Donaghey, Sara 等

'. . . and each slow dusk a drawing-down of blinds'

WILFRED OWEN Anthem for doomed youth 1917

La Milpa is a medium-sized Maya city in northwestern Belize. Discovered by the late Sir Eric Thompson in 1938, it has been investigated since 1992 by Boston University (Hammond 1991; Hammond & Bobo 1994; Hammond et al. 1996; Hammond & Tourtellot 1998; Tourtellot et al. 1993; 1994; 1996), research which has elucidated an unusual though not unique developmental trajectory, culminating in an explosive expansion of population and both monumental and domestic architecture in the period AD 750-850.

La Milpa is roughly equidistant from the very large cities of Tikal and Calakmul, some 90 km to the southwest and northwest respectively, about 20 km east from the major centre of Rio Azul in northeast Guatemala, and 40 km west of Lamanai, on New River Lagoon in the coastal plain of Belize (see Tourtellot et al. 1993: [ILLUSTRATION FOR FIGURE 1 OMITTED]), and covers an estimated 78 sq. km. The site core lies 190 m above sea level on a limestone ridge dissected by ravines, at 17[degrees] 50[minutes] 06[seconds] N, 89[degrees] 03[minutes] 6[seconds] W (UTM 16Q BQ 2-82-637E, 19-72929N). Monumental architecture, with buildings up to 24 m high and 90 m long, covers some 650x400 m (26 hectares), in two areas linked by a sacbe causeway spanning the narrow neck of land between the eastern and western drainages [ILLUSTRATION FOR FIGURE 1 OMITTED].

The northern sector includes the Great Plaza (Plaza A), covering nearly 20,000 sq. m and bordered on the east by three major temple-pyramids, Structures 1-3. Within it is a fourth pyramid, Str. 10, and two ball-courts, Strs. 6-7 and 11-12, unusually aligned on opposite axes (Schultz et al. 1994). Str. 10 faces south towards the 90m-long 'palace' Str. 8: their construction as an ensemble seems to have occurred late in the history of the Great Plaza, partly blocking Str. 3 and access to the raised court behind Str. 9. Strs. 9 and 2 are also axially aligned across the plaza; like most other structures there, they date in final form to the Late/Terminal Classic period between AD 750 and 850, during which numerous and rapid modifications were instituted and a number of stelae dedicated (Grube 1994). The lack of buildings on the north and northwestern sides of the plaza suggests that the final redevelopment remained incomplete when La Milpa was abandoned. Occupation deposits and plaster floors of the Late Preclassic (400 BC-AD 250) in almost every excavation to bedrock (Hammond et al. 1996) suggest an initial settlement covering at least 1.5 ha., and looters' trenches into most major buildings show a complex, though modest, construction history in the Early Classic period (AD 250-600).

At least four stelae were dedicated in the Early Classic (Grube 1994), indicating a claim to rulership by the lords of La Milpa, so this lack of architectural substance is intriguing; part of our objective in 1998 was to flesh out the bare bones of the site's Early Classic history. In 1996 a royal tomb of c. AD 450 was found close to Stela 1 and the northwest angle of Str. 1: since symmetrically positioned elite burials of this period are known from the nearby site of Rio Azul (Adams 1990), we sought its putative partner southwest of Str. 1 (Op. A23, A32). Although no tomb was found, we located a line of Early Classic cached offerings in front of the building, coeval with its buried initial phase. There continued to be a paucity of Early Classic pottery, and also of early Late Classic (Tepeu 1 equivalent) ceramics from La Milpa in general, indicating that this elite activity took place not only within a fairly modest community, but in one that rapidly faded into insignificance for at least two centuries. We believe that this decline, and the subsequent repopulation and florescence of La Milpa in the 8th century, are linked with the wider geopolitics of the contest between Tikal and Calakmul adumbrated by Martin & Grube (1995). Unfortunately there are few monumental texts surviving, with the ruler Ukay's Stela 7 of 9.17.10.0.0 (30 November 30 AD 780) having the only completely legible inscription at La Milpa, but the enormous amount of Late Calssic construction in both the site core and its suburbs make it second only to Caracol in size among Maya sites in what is now Belize.

Among the buildings of the florescence was Str. 4, a long, low range north of Str. 1, studied for the first time in 1998. It was decorated with elaborately modelled relief polychrome stucco, like the coeval Strs. 1 and 5, and it is likely that the uninvestigated Strs. 2 and 3 also had decorated facades: the eastern side of the Great Plaza must have been a colourfully impressive sight. Evidence that all these buildings had been abandoned before AD 900 came from area excavation of Str. 86, a dwelling incongrously located within the plaza and unlikely to have been there while it still functioned as a formal space: associated pottery fell within the Tepeu II/III tradition, with no trace of Postclassic types. Two nearby residential groups showed evidence for some post-urban occupation, however: the Pl. 183 Group yielded a small side-notched chert arrowhead of Protohistoric date, and the Pl. 76 Group, apparently abandoned unfinished in the 9th century, exhibited new rubble construction over the humus developed on the original walls; a Protohistoric date is arguable. There was also a curious episode of monument resetting, apparently around the time of the Spanish conquest (Hammond & Bobo 1994).

On the basis of substantial investigation from 1992 to 1998, the Great Plaza area thus seems to have been initially settled in the Late Pre-classic c. 300 BC, had a modest prosperity in the Early Classic, a Late/Terminal Classic apogee, and only intermittent occupation thereafter; to expand this understanding to the hitherto little-investigated southern sector of the site core, especially the massive southern acropolis south of Plaza C which seems to have been the royal palace, some 70% of the 1998 operations took place there. Excavations were also carried out in the Pl. 151 Group to the southeast, one of several putative elite residences encircling the acropolis, to complement those of 1996 in the Str. 69 Group to the west, which had demonstrated a complex but unexpected sequence including a deliberately buried building with painted floors and walls still standing several metres high (Hammond et al. 1996: 87-8).

Plazas B and C were known to be of Late/Terminal Classic date with little antecedent occupation, although here as elsewhere at La Milpa there is sparse Late Preclassic pottery: further trenches behind the one major pyramid in this sector, Str. 21, and careful study of its architecture now lead us to suggest that far from being early, as initially surmised from its rough external masonry, Str. 21 was never completed, lacking both a stair and a superstructure platform. Just north of Plaza B, adjacent to Str. 43, a quarry abandoned with large limestone blocks prised from the workface and a stockpile of construction material may have been part of this work programme: other evidence suggests that the entire area between Plazas A and B was in the course of redevelopment when La Milpa was abandoned.

A similar phenomenon was documented at the southern end of the acropolis: the simpler layout and lower elevation of construction there, compared with the northern courtyards between Strs. 32 and 39, suggested accretional development southwards. Study of numerous looters' trenches complemented by several strategically-placed cuts showed three zones of construction: the oldest and most complex centred on the Pl. 115 courtyard and bounded on the south by Str. 38; the next, completed but with a much shorter constructional history, around the Pl. 120 court south to Str. 39; and the latest south to Str. 44. North of Str. 36, Op. B70 demonstrated three major construction phases in the Late/Terminal Classic (the last with a finely preserved masonry drain from the courtyard through the flanking building to the eastern ravine), as did Op. B73 in the Pl. 115 courtyard floor. Pl. 120 to the south had one phase of flooring, although Str. 39 on its south side was remodelled, suggesting extended use.

None of the structures south of Pl. 123 was ever completed, however: Platforms 131 and 130 had rubble fills never retained by finished walls, Pl. 129 had an incomplete fill, and the flanking areas 127 and 128 were natural hillslope with patches of exposed bedrock, enclosed by long rubble walls with a slab core (Strs. 132-134) as the initial stage of courtyard construction. Within them were lines of very large, roughly quarried limestone boulders, the framing for terraces: a quarry hollow at the south end of Area 127 had been partly infilled with rubbish, preserving even the pick marks on the vertical walls. Str. 44 proved to be an amorphous bank of rubble, piled over the razed remains of a building with a red plaster floor and masonry superstructure: the top of the bank lay on the same elevation as the base of Pl. 130, corroborating the Maya intention to construct a level Pl. 129 between them.

The overall scheme, interrupted in mid-execution by the apparently sudden abandonment of La Milpa, seems to have been to create a grandiose approach to a new south-facing throne room in Str. 39, with Pls. 129, 130, 131 and 123 rising successively northwards, perhaps with flanking buildings on the completed areas 127 and 128 to the west and east. Corroboration of this come from Op. B81, an axial trench through Str. 39, which showed that its original northward orientation, with access via a broad stair and terrace from the Pl. 120 courtyard, was reversed by blocking the doorway and constructing a new, south-facing red-painted bench; it was unfortunately not possible to expose all of this, so whether it was decorated remains unknown.

It seems to have been the last of a succession of 'seats of power' on the axis of the acropolis: three earlier benches were penetrated by a looters' trench into Str. 38, facing north on to the Pl. 115 courtyard. They were built one in front of the other, the earliest being painted in plain red specular haematite pigment, the second with a greenish (originally probably Maya Blue) plaster top and red front, and the last with a similar top and front augmented with relief decoration. Both the second and third benches had false 'throne legs', set forward, of tapering trapezoidal elevation (as on the freestanding thrones at Piedras Negras and Palenque) outlined in blue-green; the upper overhanging cornice of each had been ripped off, and scorch marks on the scar and on the floor below showed that something (probably pom, copal incense) had been burnt as part of the termination ritual for the final and most elaborate throne. This was followed by infilling of the room with a massive pier resting on the thrones and blocking the doorway, outside which yet further burning ritual took place, and which formed the base for an upper room apparently facing south; at the same time, a south-facing terrace was built on to the back wall of Str. 38, effectively reversing its orientation towards the newly built Pl. 120 courtyard and the new throne room in Str. 39.

A striking repetition of this ceremonious throne-room construction and destruction was uncovered in Str. 65, a putative elite residence set on a private courtyard east of the acropolis, with a C-shaped plan of main block and flanking wings, approached by a stair and terrace. The central block consisted of two rooms: the floors were painted in deep red specular haematite, the walls in a light red with darker red framing the doorways and in a dado along the base. The inner doorway was narrower than the outer, focussing attention on the polychrome front of the bench in the inner room. The bench was 4 m long, 1.4 m deep and 0.65 m high [ILLUSTRATION FOR FIGURE 2 OMITTED]: only the central portion was visible from outside, and two relief false legs of modelled stucco were attached to the front to create the impression of a freestanding throne, painted in blue, light red and deep red specular haematite. Raised balusters at each end of the bench had been removed, leaving a scar on the plaster: one block was found in the room infill, which was clean and tightly packed, without air-spaces or soil, suggesting deliberate demolition and burial of the axial chambers.

The western lateral chamber, on the other hand, seems to have collapsed naturally after abandonment: two granite metate fragments found on its floor, one inverted over a crystalline limestone handstone with traces of red pigment, the other over a granite mono fragment, together with a pigment-stained handstone of metamorphic rock, may well have been used for preparing and polishing the haematite for the floors and walls. This chamber had its doorway narrowed at some point in its history, and although the central range has only a single phase of construction and use, excavations in the Pl. 68 Group, 25 m south, suggest this was an extended period of activity, with deposition of a dense mass of pottery (including elaborate polychromes - the first, and so far only, substantial body of such ceramics at La Milpa) arguably discarded by the inhabitants of Str. 65, as subfloor fill.

The entire history of these detached groups, as for the southern acropolis as a whole, falls within the Late/Terminal Classic period of the 8th and early 9th centuries, matching what we know of the major constructions around the Great Plaza but lacking the antecedent Late Preclassic and Early Classic occupation of that area, clearly the initial nucleus of the La Milpa community.

Outside the site core, survey and settlement excavations continued the research strategy established in 1992 to study that community. On the East Transect (ET), extended to 6.5 km in 1996, Gloria Everson (Tulane University) completed testing of 28 randomly-selected residential groups, disproving an earlier hypothesis that the huge late population growth at La Milpa (to an estimated 46,000 within a 5-km radius) was housed in perishable, more transient

quarters than usual for the Classic Maya. People lived in houses with plaster floors, and at least footing walls (often with low interior benches as well), and some had vaulted roofs; while in one case a late vaulted building appears to rest on an Early Classic platform, everything else was again Late/Terminal Classic, dated by dense middens packed tightly against exterior house walls and buried under subsequent wall tumble. Clear quantitative distinctions can be drawn between the range of domestic pottery in these suburbs and that used in the elite dwellings of the site core (Kerry Sagebiel pers. comm.)

Investigation of the minor centre o f La Milpa East (LME), 3.5 km east of Plaza A on the ET, showed that three sides of the third-largest plaza at La Milpa have long range-type structures, in most unusual inter-cardinal positions, with a small 5-m temple mound on the eastern side. In front of one corner of this pyramid the plain Stele 19 was erected, an altar in front, and incensario fragments around it suggesting Terminal Classic veneration. We are not sure that this minor centre was residential rather than ceremonial: its formality, large size, and hilltop siting in view of Plaza A suggest a middle-level administrative centre with responsibility for a surrounding area some 1.5 km in radius, probably one of several.

The 1998 South Transect (ST) extended to 3.8 km south of site datum, and for speed was 260 m wide (rather than 500 m as on the NT and most of the ET), producing 0.66 sq. km of new map: results were gratifyingly similar to those on the other transects, despite encountering a higher proportion of bajo (swamp) terrain and a more rugged landscape. The ST crossed four ridges, 10-23 m high, with 600 m of seasonally-flooded and snake-infested terrain in the middle, and ended short of a large bajo which we suspect defines the southern border of La Milpa.

The 456 mapped ST features include, besides the usual house-size mounds and basal platforms, smaller platforms, rock piles, terraces and stoney berms, with the great majority of the non-residential features on the lower ground. Only 131 (29%) of the features were residences, an extraordinarily low percentage for Maya settlements, but due to the very high numbers of utilitarian features (terraces and berms) rarely recorded elsewhere. The percentage of houses varies from 56% on the ridge crest at the north end, to 16% at the south end of the transect; assuming 4 persons/house gives the equivalent of 794 people/sq. km, 7% higher than population densities calculated for previously-mapped parts of La Milpa. No excavations were carried out, but we would expect a Late/Terminal Classic date for most or all structures.

The most distinctive group is located on a small hill at N2725, with three main structures, the tallest on the east, flanking an isolated, elevated, and open-cornered plaza 3-4 km south of Plaza A, approximately the same distance as LME lies east, although less architecturally impressive. Tourtellot et al. (1998) proposed an explanation for this spacing based on demographic and administrative convenience for a middle level managerial centre.

Overall, the 1998 season both confirmed and expanded on our previous conclusions about the history of La Milpa: the limits of the community and the tactical use of terrain within it, creating an engineered landscape in pursuit of adequate subsistence and in the face of rapid Late/Terminal Classic population growth, are now established. The nucleated Late Preclassic community may have spread out slightly more than we surmised, but in spite of the presence of at least four Early Classic stelae (of which only Stele 10 is in situ) there is still little evidence for architectural aggrandisement, and much for a dramatic decline in La Milpa's fortunes in the 6th and 7th centuries AD. What is certain is that coeval with the resurgence of Tikal after AD 693, La Milpa recovered, and in the 8th century underwent rapid urban development, including the construction of the entire southern sector of the site core and the subsequent complex remodelling of the acropolis palaces.

One of the principal revelations of the 1998 season is the suddenness of the end, most dramatically demonstrated at the southern acropolis, where a major public works programme under royal patronage was abandoned unfinished, but notable also in the incompletion of the Str. 21 pyramid and the north side of the Great Plaza. There was no slow dusk and drawing-down of blinds at La Milpa: the Maya collapse here came quickly, and our current model of it as an extended process will, once again, require adjustment.

Acknowledgements. Funding was provided by the National Geographic Society (Grant #6112-98), Boston University, Raymond & Beverly Sackler and an anonymous donor. Permission to work at La Milpa was granted by the Government of Belize (through John Morris, Archaeological Commissioner) and the landowner, Programme for Belize, which also provided logistic help through Joy Grant and Ramon Pacheco. Apart from the authors, survey and excavation work was supervised by Jeremy Bauer, Ryan Mongelluzzo, Ben Thomas and Marc Wolf (Boston University), Jason Gonzalez (Southern Illinois University), Julie Kunen (University of Arizona) and Kristen Gardella. Pottery was analysed by Kerry Sagebiel (University of Arizona) and Laura Kosakowsky (Boston University), and human burials by Julie M. Saul (Lucas County Coroner's Office, Ohio). Site plans were drawn by Jan Morrison and finds by Candida Lonsdale. The field laboratory was directed by Cynthia Pinkston (University of Maryland) and the field camp by Helen Warren (Pennsylvania State University). Much of the work was carried out by undergraduates from Boston University, New York University, Drew University, Bates College and Oxford University, and by our Belizean team from San Felipe Village. We are grateful to our colleagues Jenny Bacon, Chantal Esquivias, Elizabeth Graham, Gigi Green, Mark Hodges, Sheena Howarth, John Masson, Duncan Pring, Rick Russo, Vernon Scarborough, George Stuart, Fred Valdez, Jr and Bryan Woodye for their help in various ways.

References

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

GRUBE, N. 1994. A preliminary report on the monuments and inscriptions of La Milpa, Orange Walk, Belize, (new series) 42: 217-38.

HAMMOND, N. 1991 The discovery of La Milpa, Mexicon 13: 46-51.

HAMMOND. N. & M.R. BOBO. 1994. Pilgrimage's last mile: late Maya monument veneration at La Milpa. Belize, World Archaeology 26: 19-34.

HAMMOND, N. & G. TOURTELLOT. 1998. La Milpa: a Classic Maya city of the eastern lowlands, Proceedings of the 13th UISPP, Forli 1996 5(17): 493-8.

HAMMOND, N., G. TOURTELLOT, S. DONAGHEY & A CLARKE. 1996. Survey and excavation at La Milpa, Belizo, 1996, Mexicon 18: 86-91.

MARTIN, S. & N. GRUBE. 1995 Maya super-states, Archaeology 48(6): 41-6

SCHULTZ, K.C, J.J. GONZALEZ & N. HAMMOND. 1994. Classic Maya ballcourts at La Milpa, Belize, Ancient Mesoamerica 5: 45-53.

TOURTELLOT, G., A. CLARKE & INI. HAMMOND. 1993. Mapping La Milpa: a Maya city in northwestern Belize, Antiquity 67: 96-108.

TOURTELLOT, G., G. EVERSON, L. KOSAKOWSKY & N. HAMMOND. 1998. Suburban administration: minor centers at La Milpa, Belize. Paper presented in symposium: The social implications of ancient Maya rural complexity, 63rd Annual Meeting of the Society for American Archaeology, Seattle (WA).

TOURTELLOT. G., J.J. ROSE, N. GRUBE, S. DONAGHEY & N. HAMMOND. 1994. More light on La Milpa: Maya settlement archaeology in northwestern Belize, Mexicon 16: 119-24.

TOURTELLOT, G., J.J. ROSE & N. HAMMOND. 1996. Maya settlement survey at La Milpa, Belize, 1994, Mexicon 18: 8-11.