Chiik Nah was originally located during the 1984 survey of a 250 meter by 10 km transect between the center of El Pilar and the Belize River (Figure 1).  At that time the mound was designated as 272 (the transect number)-32 (the site number).  We have christened it Chiik Nah, a Maya phrase meaning “house of the Coati” in honor of an abandoned pet that was befriended by the El Pilar site guards and frequented the excavation area during the early part of the excavation season until it mysteriously disappeared. The mound at Chiik Nah is about 0.5 meters high and about 5.5 meters in diameter.   For excavation and mapping purposes, we included a surround of about 10 meters in all directions from the center of the mound in the site definition.Chiik Nah was partially cleared with small trees and shrubs removed. Larger trees were left standing, in order to coincide with the project’s effort to preserve the forest environment. Using an electronic distance meter (EDM) and back sights from EP12, a previously established control point in the parking lot to the south of Tzunu’un, two new control points (temporarily designated 32CP 4 and 32CP5) were set up within the site limits and marked with cement posts.  These control points were then used to set up a 4m north-south grid over the site.  Posthole tests were conducted along the grid to determine artifact distribution throughout the site.  No postholes were dug on the house mound itself to avoid damaging possible architectural features.  Posthole tests were dug to bedrock with posthole diggers and spades.  All excavated soil was screened through 1/4-inch mesh. Based on the data collected from the posthole tests, two 2m by 2m test trenches (TT1 and TT2) were opened in areas near but not on the mound with relatively high artifact counts.   TT1 and TT2 were respectively located to the northeast and to the southwest of the mound.  Later, 2m by 2m test trenches (TT3 and TT4) were opened on the mound itself (Figure 3).  TT3 was located on the southeast corner of the mound to define the mound’s edge.  TT4 was completely located on the mound, directly north of TT3.  A 1m balk separated TT3 and TT4.  All test trenches were set up on even UTMs to facilitate mapping.Chiik Nah was partially cleared with small trees and shrubs removed. Larger trees were left standing, in order to coincide with the project’s effort to preserve the forest environment. Using an electronic distance meter (EDM) and back sights from EP12, a previously established control point in the parking lot to the south of Tzunu’un, two new control points (temporarily designated 32CP 4 and 32CP5) were set up within the site limits and marked with cement posts.  These control points were then used to set up a 4m north-south grid over the site. Posthole tests were conducted along the grid to determine artifact distribution throughout the site.  No postholes were dug on the house mound itself to avoid damaging possible architectural features.  Posthole tests were dug to bedrock with posthole diggers and spades.  All excavated soil was screened through 1/4-inch mesh. Based on the data collected from the posthole tests, two 2m by 2m test trenches (TT1 and TT2) were opened in areas near but not on the mound with relatively high artifact counts.   TT1 and TT2 were respectively located to the northeast and to the southwest of the mound.  Later, 2m by 2m test trenches (TT3 and TT4) were opened on the mound itself (Figure 3).  TT3 was located on the southeast corner of the mound to define the mound’s edge.  TT4 was completely located on the mound, directly north of TT3.  A 1m balk separated TT3 and TT4.  All test trenches were set up on even UTMs to facilitate mapping.

Test trenches were dug by teams of two to three people using trowels and hand picks.  Hand picks were necessary due to the very hard clay loam that was encountered during excavation.  Test trenches were dug down to bedrock according to natural strata, which were further subdivided into 10 cm arbitrary levels.  Chert and obsidian stone tools, chert cores, and large potsherds found in situ were point provenienced using an EDM.  Levels and architectural features such as floors were also mapped using the EDM.  All features, levels and point-provenienced artifacts were further recorded on sketch maps.

Fill was screened using 1/4-inch mesh, but due to slow and careful digging more artifacts were found in situ than in the screen.  In addition, a soil sample from TT1 was water screened.  Only a few small flakes were found, indicating that our standard 1/4 inch screening was sufficient, and it was extremely difficult to process the clay fill through a fine screen, so no further water screening was done.  Samples of soil and plaster from cultural surfaces found in TT2, TT3 and TT4 were collected for pollen testing.  

Site 272-32 was used as an agricultural field, within the last 10 years.   Charcoal from burning the field burning was still evident, but was not collected.  No charcoal was found in undisturbed contexts. One simple bucket-and-sieve flotation was performed on a sample of fill from TT4.  There were many fresh rootlets, but no useful carbonized plant remains. Bone and other organic material was apparently not preserved in the mound. Only lithics and potsherds survived both the burning and the humid tropical environment.  Many of the lithics collected showed evidence of burning such as spalling and heat fracture.  Also, some of the wear on the lithics may have come from historic disturbance rather than prehistoric use. The potsherds collected did not appear damaged by burning, but were heavily corroded and soft. The heavily corroded appearance may have been caused by the dissolution of limestone temper. The basic soil conditions resulted in a thick patina on the lithics and may have affected the preservation of other artifacts as well. 

Stratigraphy and Temporal Sequence

Test Trenches 1 and 2 were located to investigate surfaces around the mound, while 3 and 4 were situated to explore the contents of the mound.  No surfaces were identified in TT1, but an irregular cobble surface was found in TT2, and TTs 3 and 4 each exposed a plaster floor. Unfortunately, the relationship between the floor surfaces in each trench is not clear, and we did not have time to excavate the areas between the trenches in hopes of clarification.

            Test Trenches 1 and 2: Test Trench 1, located northeast of the mound, contained few artifacts, only 174 sherds and 374 lithics, including both tools and debitage. Under the upper few centimeters of humus, the soil became increasingly clayey and hard, with many roots and small limestone rocks. The sterile substrate, composed of decaying limestone, began about 30 cm from the surface.

Test Trench 2 was 5 m southwest of the south edge of the mound. The basic profile was very similar to that in Test Trench 1.  Artifact density was even sparse, with only sherds and 191 lithics. The decaying limestone bedrock began at about the same depth as in TT1; however, in the northern half and more patchily over the rest of the test trench, lying directly on the sterile substrate, was a discontinuous expanse of tightly packed potato sized limestone lumps. This surface was unplastered, but appeared to have been prepared. 

            Test Trench 3: Test Trench 3  (Figures 4 and 5) was on the south edge of the mound. Beneath the entire excavation was the same sterile subsoil (104): decaying yellow limestone with clay, which after about 10-15 cm gives way to more solid soft limestone.

            Outside the mound area, the fill above sterile 104 is yellowish clay loam with rocks, becoming rockier and darker higher up. Stratum 113 was excavated as this fill south of the floor 110/114, with Stratum 108  as gravelly fill over most of the unit and over the floor area, and 107 as darker fill around the apparent edge of the mound before the floor was exposed. Stratum 107 is equivalent to 106, 105, and 102 in the northern part of the TT, over the mound, and no visible distinction can be made.         

            Under the mound area in the northern part of the test trench, a layer of tightly packed limestone cobbles and lumps in clay (114) were laid directly on the sterile 104, which may have been artificially flattened first.  This 114 apparently served as a substrate for a plaster floor (110), which was about 2-4 cm thick, with a well smoothed, level surface. It survived over the NW 1/3 of the test trench, and although there was no apparent border of larger rocks, the arc of the edge is close to what might be expected from the surface appearance of the mound, and we believe it is the edge as constructed.  Above the floor was the gravelly fill 108, which appeared continuous with material outside the boundary of the mound. Above this were fill units designated 106, 105 and 102, which in retrospect seem largely undistinguishable, except that they become increasingly dark and humic nearer the surface. Stratum 101 was the mound surface.

            Test Trench 4: Test Trench 4 (Figures 4 and 6) was fully within the apparent bounds of the mound, but the floor surface we encountered was only in the SW 1/3 of the trench.  As in Test Trench 3, the sterile 104 was remarkably flat, possibly artificially flattened. In the SW corner, a rocky clay fill (119) lay directly on this. It was thicker, looser, and much less organized or packed than the possibly equivalent 114 under floor 110 in TT3. In the extreme corner, there was a substantial pocket of completely loose rock, with very little clay matrix. We designated it Stratum 116, but eventually felt that it represented an area of root or burrow disturbance of 119. Above 119 was a plaster surface, or rather two. Stratum 115 was a poorly surfaced yellow plaster some 4 cm thick. In the corner, some 25 cm of Stratum 119 intervened between plaster 115 and sterile 104, but the 115 plaster formed a sloping arc, poorly preserved, that came down to the sterile 104 in the middle of the test trench. Plaster remnants rested directly on sterile 104 or with no more than a cm or two of clay intervening. Unfortunately, the plaster was only traceable in patches, never with good surfaces. While cleaning the south profile, we cut a strip 10 cm wide about 5 cm deep into the 104, and exposed a dark round stain 10 cm in diameter right against the profile, and just outside the arc of plaster 115. Excavating the stain carefully with fingers (fill was wet clay), recovered no artifacts, but produced a cone-shaped hole about 10 cm deep. The form and position suggest a posthole. Further scraping of the sterile 104 along the contact with plaster 115 failed to reveal any other similar features.

Directly on top of 115 in the corner of the trench was 111, another plaster, 4 cm thick, much whiter than 115 and with a nicely smoothed, level surface. It appeared that small discontinuous bits of this plaster extended at the same level over the sloping part of 115, at least as seen in the profiles to the south and west.  Our initial interpretation was that 111 represented a replastering of the 115 floor, either as repair, or a later occupation. Alternatively, 115 could be a substrate prepared for the finished floor 111, although it is not clear why 115 should slope down and 111 appears to continue more horizontally. Possibly at the edges of the floor we are seeing a surface marked by plaster bits from both plasters, and the sloping 115 also is eroded remnants of both, either forming a sloping edge as constructed, or slumped and eroded.

            We designated the fill over floor 111 as Stratum 109. Above 109 some fill was designated 108, essentially equivalent to that in TT3, but later we felt that we had been saw a transition between 109 and 106, the upper fill of the mound, which was distinctively humic and rocky, darker and looser than fill below. Stratum 106 was recognized over the whole trench, and the very top fill over the whole unit was 102 or 105, but there is no real distinction among them.

            Outside the area marked by the extent of plaster 115, the fill over sterile 104 was labeled 117, and is essentially similar to 113 in TT3.  Above that we recognized 112, but this now appears arbitrary, and 112 and 109 are essentially continuous.

Stratigraphic Interpretation of the Mound

            The big unanswered question is the relationship between the floors in TT3 and TT4. The top of the 110 floor in TT3 is only about 10 cm below the top of the plaster 115 in TT4, and 14 cm below the plaster of 111. However, the plaster of 110 is white like the plaster of 111, not the yellower plaster of 115, and although the absolute difference in elevation is not great, the surfaces of both 110 and 111 were flat over the extent observed, making it unlikely that there is a slope concealed in the balk. Additionally, the cobble substrate 114 under 110 is thinner and much more organized than the 119 fill under 115, and the edge of  floor in TT3 shows none of the slope of 115 in TT4.  Presumably this means that somewhere in the balk, one floor must cut the other, or form a continuous surface. The fact that in both test trenches the apparent floor edges form rather tight arcs, suggests that several small floors might be incorporated into one smallish mound.

            In any case the mound began as a floor or floors built on a clean sterile surface, and covered by a build-up of clay loam and limestone rock, with midden debris as well. The material over the floors presumably derives from wall structures; they could have been stone and clay in a wooden crib framework.  The refuse component is fairly sparse; most of the sherds are small and isolated. The lithic material suggests that debitage and tools were incorporated in fill or structures rather than produced on the spot. There is little small debitage, and nothing to suggest that any of the artifacts are in situ on an interpretable living surface

Ceramic Analysis

            Methods of Analysis All potsherds larger in diameter than a quarter were collected and washed.  In the lab, potsherds were initially sorted by presence or absence of ash temper.  Ash tempered and non-ash tempered sherds were further sorted according to the vessel part represented: rim, neck, body, base or foot.  Body sherds with decorations or waxy slip were separated into their own category.  The number and weight of sherds in each category were recorded for every excavation level.

The next stage of analysis focused on potentially diagnostic sherds.  These sherds included rims, necks, decorated and waxy body sherds, bases, flanges and feet.  All rim sherds and other diagnostic sherds were assigned a unique number and information was recorded on the following characteristics:

1. Vessel Part: Rim, neck, body, base, flange, feet or indeterminate vessel form.

2.  Slip: Presence of slip on the exterior, interior, or both exterior and interior.

3. Decoration: Incised designs, fingernail impressions, appliqué, or painting.  
4. Paste: Coarse, fine, Mars Orange, or ash paste.

5.  Vessel Form: Jar, bowl, plate, vase, or indeterminate. 

6. Diameter: Rim diameters in cm, when possible.

7.  Period: Sherds were classified according to time period (i.e. Early Classic, Late Classic, etc.), based on BRASS type collections and spot checks by Anabel Ford. 

            In general the sherds recovered at Chiik Nah were small and in poor condition.  What were initially seen as good-sized sherds in the ground often broke into tiny fragments, even when excavated with a brush, or threatened to dissolve during washing.  Because of the limited scope of excavations and poor preservation of sherds at the site there were few potentially diagnostic sherds.  A total of 264 potentially diagnostic fragments, primarily rims, were examined, but less than half of these were assigned a temporal designation.  Preliminary analyses of the sherds revealed occupation in the area from the Late Pre-Classic through Late Classic. Heaviest occupation appears to have been during the Early and Late Classic.  The Terminal Classic sherd was encountered on the surface and is the traditional pie-crust jar found throughout the central Maya region.

PERIOD AS DETERMINED BY CERAMICS

Most of the material came from fill above the floors, and there is no obvious reason to subdivide the assemblage temporally. Most of the material derives from deposits above the floors, both because there was much more fill there, and because artifacts in general were sparse under the floors. About 15% of the 240 tools showed signs of burning, and about 23% of the flake assemblage.  Almost 75% of the burned tools were from the upper fill on the mound, which also had a higher percentage of burning (21%) than other groups of strata both lower, and off the mound. This suggests to us that much of the burning is from the action of field fires on exposed mound material; however, it could also be argued that the architectural fill used on the mound suffered more prehistoric burning than surrounding materials.

            The assemblage is dominated by simple core flaking, with only about 14% of the flakes appearing to result from biface manufacture. Most of the flakes are of a size to be usable as tools; there are relatively few small flakes and even fewer really large ones, which along with the scarcity of cores, suggests that little knapping was done on site.

The contrast with the LDF Chert Site, where most of the assemblage was recognizably biface thinning flakes and small flakes and fragments were common, is striking.

Tools

Tools from all the excavated units are included in the analysis, although most came from Test Trenches 3 and 4 (see Figure 7 for a map of the locations of tools with point proveniences).  Information recorded for tools included type, weight, dimensions, blank form, and presence or absence of burning.  Burning was identified by the presence of heat spalls, pot-lid fractures, crazing, and discoloration.  The following blank forms upon which tools were made were distinguished: Hard hammer flakes, Biface thinning flakes, Indeterminate flakes, Blades, Macroblades, and Other.  The “other” classification was usually used for tools made directly from nodules or fragments of nodules.  In one case a tool was made on a piece of a broken axe and the classification “axe” was used for its blank form.

1.  Scraper: These are arbitrarily distinguished from the large number of flakes with minor retouch or edge damage by requiring them to have more than 2 cm of steep, orderly retouch on at least one edge.

2.  Notch: Notches have an apparently purposeful creation of a notched scraping edge.  Often gravers (below) were made by removing two notches to form a spur, which at times made it difficult to distinguish which property was desired, the notch or the spur.

3.  Tool Edge: This classification is given to tools that possess an edge of some sort, but are too irregular to classify as a chopper, scraper or graver.  They are usually flakes with a short length of apparently purposeful steep retouch.

5.      Multiple Tool: These are tools possessing several small irregular used edges.

6.      Flake Graver: These consist of small, mostly unifacial spurs produced in a number of ways on small to medium sized chert flakes.  The classification of flake graver has priority over classifications multiple tool and notch, as these spurs were sometimes included on the end of flakes possessing several irregular retouched edges, and were often formed by the removal of two large notches from either side of the intended spur.  Flake gravers are found on a wide variety of flakes in a variety of sizes and forms, making it difficult to describe the “typical” flake graver. 

7.      Shaped Graver: These distinctive tool forms are characterized by small points on a flake worked around the entire circumference with steep, c. 90 degree, unifacial retouch.  This tool is of a size held comfortably between the thumb and forefinger or hafted in a socket, and other analysts have called these drills.  We make no classificatory distinction between them, but several different forms appear consistently. Most are a “tear drop” or “corn kernel” form , but some have one or both shoulders sharply defined below the point. Most have well knapped shoulders indicating that it was a deliberate feature the tool.  Of final note is their uniformity in size.  Most are within a few millimeters of each other in length and width.  The typical shaped graver is about 22mm long, 10mm wide, and 3mm thick.

8.       Blade graver: Only one specimen was found at 272-32, but in examination of the tools from 272-136 they can be recognized as a distinct tool form.  Blade gravers are long, narrow tools created by steep, unifacial retouch on a thick, narrow flake or macroblade segment.  When whole, they are c. five or six times as long as they are wide, almost as thick as they are wide, and often display heavy battering or stepping on the sides.

9.      Edge Dulled Flakes: These flakes are distinctive oval pieces, usually 3-6 cm in diameter, severely dulled with bifacial buffering around much of the perimeter.  Although relatively uniform in size and shape, edge-dulled pieces were categorized by the presence of this extensive dulling, not their oval shape. Such pieces are often called wedges.  None were found at Chiik Nah, but they were common at 272-136.

10.  Prismatic blade. These tools are classified by small narrow flakes with multiple parallel scars. These are counted as tools even when not retouched. Only obsidian examples were found at 272-32, while they appear in other collections in both chert and obsidian.

11.  Axe: These tools are ovoid, relatively thick bifaces with finished edges, often showing dulling or damage to a bit at the wider end. These are sometimes called “general utility bifaces” in some of the literature, and “Biface ovates” for Tikal.

12.  Chopper: Heavy-duty tools with a relatively straight, battered edge, usually bifacial, are classified as choppers.  Relatively unbattered choppers proved difficult to differentiate from cores in the field.

13.  Rough Biface: These bifaces exhibit the presence of irregular edges, lack of symmetry, or extreme thickness and may be considered unfinished due to their asymmetrical form.

14.  Core: Any piece of material possessing several relatively large flake removals and showing no signs of use.

15.  Prismatic Blade Core: These are small cores with multiple small blade removals. None were recorded at Chiik Nah, but they are recognized in the collections from 272-136.

16. Ground Stone: The ground stone is not included in the tool statistics.  Two fragments of manos were found on the surface of the mound outside of the two test trenches. Both were fire-cracked; one was limestone, the other granite.

Frequency of Tool Types (Chert unless noted)

Summary of Excavation Results

            The Chiik Nah residential type at El Pilar represents a common residential unit in the local area.  This small mound was the collapse of a sequence of constructions of domestic character.  None of the constructions of Chiik Nah would have been large enough to shelter more than a small family, and the collections suggest that it served general purposes. The previously excavated domestic structures in the BRASS area exhibit considerable variability in the details of construction from consolidated cobble foundations to major faced limestone buildings. Chiik Nah appears to be well within the range of extant variation with raised foundations, floor ballast fill, and plaster surfaces. The artifacts are all of a type that might be expected of a household; generalized in strategies of use more common at the lower end of the economic spectrum, despite the mound’s close proximity to the El Pilar ceremonial core.  This phenomenon of economic integration has been noted elsewhere at El Pilar and in the Maya area as well. 

At this point, it is difficult to specify the economic strategy of occupations at Chiik Nah.  The generic quality of the assemblage suggests ordinary series of activities. Tools found at the site do not point to any specialized craft activities.  Some were undoubtedly used for basic domestic activities, while others probably served as agricultural implements. 

The soils at Chiik Nah are clayey, hard to work, and not especially optimal for agriculture. On the other hand, they are not unsuitable either, as suggested by the recent evidence of maize agriculture in the area.  The occupation concentrates in the Classic Period, with little evidence of earlier and later occupation.

Little doubt, the growth of El Pilar in the vicinity influenced occupation in this zone. The survey map shows several other small mounds in the vicinity. Nearby sites in more productive soil areas evidence higher settlement density and larger residential units. This is suggestive of a pattern of land use where the best soils are occupied before other areas.  Such a pattern is consistent with regional patterns of the availability of well-drained ridges and settlement density between Northern Belize and the central Petén. Further, the pattern of occupation between Chiik Nah and Tzunu’un mirrors the pattern of land use expansion and contraction noted for the Belize River area in the surveys where the areas of clayey soil dominant in the marl based foothills were last occupied in the late Preclassic and were the first abandoned in the Terminal Classic.  It may be the pattern of occupational preference is a cross-scale pattern.  This will need to be tested.  

Chiik Nah is a single-structure unit is close to the center of the site, as are the other units in its vicinity. While the other units are unexcavated, it is not known whether they were contemporaneous to Chiik Nah, but given the BRASS survey data where 98% of the residential units had Late Classic occupation, we can assume that most had some degree of contemporaneity at least for that time period. Another possibilit