3

Lillian's Site

Mark D. Varien

Contents

Introduction

Surface Architecture (Sampling Stratum 1)
The East Roomblock Unit: Structure 2
Construction
Walls
Floor
Feature 1 (Burned Spot)
Roof
Dating
Stratigraphy
Structure 2 Interpretations
Structure 3
Construction
Walls
Floor
Roof
Dating
Stratigraphy
Structure 3 Interpretations
The West Roomblock Unit
Construction
Walls
Stratigraphy
West Roomblock Unit Interpretations
Wall Connecting the East and West Roomblock Units
Construction
Wall
Dating
Stratigraphy
Feature 3 (Burned Spot)
Structure 4
Construction
Walls
Sampling Stratum 1 Summary and Interpretations
Pit Structure and Courtyard Areas (Sampling Strata 2 and 3)
Structure 1
Construction
Walls
Feature 1 (Bench 6)
Feature 8 (Pit)
Feature 2 (Pilaster 6)
Feature 6 (Bench)
Feature 7 (Pilaster)
Floor
Feature 5 (Hearth)
Feature 3 (Deflector)
Roof
Dating
Stratigraphy
Structure 1 Interpretations
Inner Periphery (Sampling Strata 4 and 6)
Feature 1 (Pit)
Plaza (Nonstructure 3)
Construction
Walls
Inner Periphery Artifacts
Inner Periphery Summary and Interpretations
Midden (Sampling Stratum 5)
Burial 1
Midden Artifacts
Midden Interpretations
North Outer Periphery (Sampling Stratum 7)
Feature 1 (Slab Cluster)
South Outer Periphery (Sampling Stratum 8)
Feature 1 (Post Hole)

Lillian's Site Artifacts

Lillian's Site Summary and Interpretations


Introduction

Lillian's Site, 5MT3936, was initially recorded in 1986 as CC86-104 by a Crow Canyon survey team (Van West et al. 1987). It was test-excavated in 1988 by Crow Canyon as part of the Sand Canyon Archaeological Project Site Testing Program. The site rests on a mesa top at 2073 m (6799 ft) and lies 1.7 km north-northwest of Sand Canyon Pueblo (Figure 1.7). A small tributary of Yellow Jacket Canyon is the closest drainage to Lillian's Site; it is located 300 m to the east of the site and empties into Yellow Jacket Canyon 3 km to the north.

Dense sagebrush and scattered pinyon and juniper are the most common plants growing on the site, with numerous other plants present as well (Figure 3.1 and Figure 3.2). The area around the site has been used as rangeland during the historic period, and the mesa top was chained in the 1960s to try to improve the quality of the range. Grazing has been the only historic-period land use.

Surface remains at Lillian's Site include a masonry rubble mound, a pit structure depression, a fallen masonry tower, and a midden (Figure 3.3). East and west of the pit structure depression lie areas referred to as the "inner periphery." The inner periphery is characterized by a low (ca. 20 cm) topographic rise, areas of light rubble scatter, and a moderately dense artifact scatter (lower artifact density than the midden, but greater density than in other areas of the site). Surface remains did not permit specific interpretation of the inner periphery, although the presence of features and nonmasonry architectural units was anticipated. An "outer periphery" surrounded the more clearly defined cultural units. The outer periphery is characterized by a light artifact scatter and no surface evidence of architecture or features.

The surface remains were used to define sampling strata, provenienced in the field as arbitrary units (Figure 3.4). Sampling Stratum 1 consists of the well-defined surface rubble mounds, including both the masonry roomblock and tower. The boundaries of this sampling stratum were drawn where the outside walls of the roomblock and tower were anticipated, and not around the edge of the rubble scatter. The roomblock rubble mound rises 65 cm; the tower rubble is 1 m high.

The pit structure depression defines Sampling Stratum 2; this depression is approximately 40 cm deep. Sampling Stratum 3, the area surrounding the pit structure depression, is bounded by the masonry roomblock on the north, the tower rubble on the south, and the inner periphery on the east and west. Sampling Stratum 3 is interpreted as a small courtyard associated with these architectural units.

Sampling Strata 4 and 6 make up the inner periphery. The boundaries of Sampling Stratum 4 were drawn around an amorphous scatter of sandstone rubble. This rubble is not as dense or as topographically high as the Sampling Stratum 1 rubble. Sampling Stratum 6 is characterized by a low topographic rise and moderate artifact density, but almost no sandstone rubble.

Sampling Stratum 5 equates to the midden (provenienced as Nonstructure 1). The midden boundary is sharp and is drawn on the basis of a high artifact density and an earth mound 60 cm high.

The outer periphery consists of Sampling Strata 7 and 8. Both sampling strata extend to the limit of the light artifact scatter that is associated with the site. Sampling Stratum 7 is located on the northern end of the site, surrounding the architectural units. Sampling Stratum 8 surrounds the midden on the southern end of the site.

Thirty-five randomly selected sampling units and five judgmentally selected test pits were excavated at Lillian's Site (Figure 3.4). The excavation of these units resulted in the documentation of the cultural features illustrated in Figure 3.5. These include two masonry roomblocks connected by a low masonry wall (a masonry room, Structure 2, was found in the east roomblock), a masonry-lined pit structure or kiva (Structure 1), a masonry tower (Structure 4), the midden (Nonstructure 1), the courtyard (Nonstructure 2), and a plaza area partly enclosed by a low, circular, masonry wall (Nonstructure 3). These features are associated with the Pueblo III occupation of the site, which is interpreted as dating to the early A.D. 1200s. A post-and-adobe roomblock, found below the eastern masonry roomblock, was also discovered during testing (Structure 3 is a post-and-adobe room that is a part of this roomblock). This post-and-adobe roomblock is believed to be associated with a Pueblo II occupation that dates to the late A.D. 1000s.

The grid at Lillian's Site was oriented along what were interpreted as wall alignments in the roomblock. Subsequent excavations demonstrated that the alignments interpreted as walls were, in fact, wall fall. Excavation exposed two separate roomblock units, termed the east and west roomblock units, and the grid was not oriented on either. Grid north on Lillian's Site is 28 degrees 30 minutes west of true north. Vertical Datum A, arbitrarily assigned an elevation of 100 m, is located on top of the rebar set in concrete at 92N/100E. Three other site datums were established at 100N/88E, 100N/100E, and 100N/108E.

The following excavation descriptions are organized by sampling stratum. Cultural study units discovered within a sampling unit are discussed first, followed by descriptions of sampling units in which no structures were found. Sampling stratum discussions include descriptions of both probability sampling units and judgmentally located sampling units.

Surface Architecture (Sampling Stratum 1)

Sampling Stratum 1 encompasses the masonry surface architecture, including the roomblocks and the tower. One hundred thirteen sampling units are present in Sampling Stratum 1. Although as a general rule we tried to excavate four sampling units in each sampling stratum, only two (104N/99E and 102N/95E) were excavated in Sampling Stratum 1 because the remaining two were noncontiguous squares inside the tower. Expanded excavations would have been needed to make the excavation of the randomly selected sampling units logistically feasible, and excavating large portions of structures was not within the scope of the Site Testing Program.

The random sampling (or "probability") units in Sampling Stratum 1 landed on walls. The excavation of these units required expanding them with judgmentally placed 1-×-1-m units located to the north and south (Figure 3.4 and Figure 3.5). Judgmental units excavated in Sampling Stratum 1 include 103N/99E, 105N/99E, 101N/95E, and 103N/95E.

Nonrandom excavations in Sampling Stratum 1 also included the more complete delineation of walls exposed by the probability sampling units. The tops of the walls were followed by sweeping and troweling to determine the architectural limits and layout of the masonry roomblocks. A compromise was struck between the goals of data recovery and minimizing the impact to the site. Poorly preserved walls that could not be followed by simply sweeping and troweling were not identified. The attempt to define walls was most successful on the outside of the roomblocks, but less successful on deeply buried cross-walls inside the roomblocks. Wall trenching did allow identification of separate east and west roomblock units and a curved wall that connects the two.

Sampling Stratum 1 excavations exposed portions of a masonry room, Structure 2. A post-and-adobe room, Structure 3, was found beneath Structure 2. These rooms are discussed first, followed by a discussion of the judgmental excavations in the roomblocks. The tower, Structure 4, is discussed last. Structure 4 is located within Sampling Stratum 1. No Sampling Stratum 1 excavations took place inside Structure 4, but two Sampling Stratum 4 sampling units were excavated just outside of this tower and allow us to describe a portion of the exterior walls.

The East Roomblock Unit: Structure 2

Structure 2 is a masonry surface room discovered in the 1-×-1-m probability sampling unit 104N/99E, and this masonry room is a part of the east roomblock unit at Lillian's Site (Figure 3.6, Figure 3.7, and Figure 3.8). The probability sampling unit fell over the northwest corner of the room, so that the sampling unit lay both inside and outside Structure 2. A judgmentally located 1-×-1-m unit, 103N/99E, was also excavated to provide an exposure suitable for stratigraphic interpretation of Structure 2. An earlier room, Structure 3, was found beneath Structure 2 in these excavation units.

A second judgmentally located square was placed on the north side of the probability unit to interpret the contexts that fell outside Structure 2. Wall trenching indicates that Structure 2 dimensions are approximately 5.2 m by 2 m; however, there may be internal cross-walls that were not located during wall outlining, in which case, this estimate may be too large. In addition, the south wall was poorly preserved. More complete excavation is needed to confidently determine Structure 2 dimensions.

Construction

        Walls. Structure 2 excavations exposed small portions of the north and west walls. The north wall is 58 cm high and 33 cm thick; the west wall is 54 cm high and 31 cm thick. Walls are semicoursed and constructed with sandstone blocks. Both the north and west walls are double-stone wide in cross section. Smaller stones were used on the interior face, and larger stones on the exterior face. The northwest corner of Structure 2 is bonded.

Little shaping of the stone faces is present on the interior of Structure 2, although some pecking and flaking was noted. Extensive stone shaping characterizes the exterior of the north wall; over 95 percent of the exposed portion of this wall is pecked-block masonry. Joints between masonry courses measure 1 to 2 cm, and chinking occurs in approximately 15 percent of these joints. Mortar is a brown to red brown silty clay loam, similar in both color and texture to the undisturbed B horizon on the site. Footings are absent on the north and west walls; instead, these walls rest on the fill of an earlier room, Structure 3.

        Floor. The Structure 2 floor (Figure 3.8) is assumed to be a use-compacted surface; however, the floor could not be confidently identified during excavation. The Structure 2 floor was expected somewhere beneath the Structure 2 wall fall, near the base of the Structure 2 walls. Wall fall completely filled Structure 2 down to the base of the walls. Immediately beneath the Structure 2 wall fall lay the burned structural collapse of the room below, Structure 3.

The poor definition of the Structure 2 floor can be accounted for by rodent disturbance, the collapse of the structure walls directly onto the floor, and the fact that the floor is simply the use-compacted fill of an earlier structure. No artifacts could be associated with the Structure 2 floor, because of the poor preservation.

        Feature 1 (Burned Spot): A single feature, a burned spot, was located in the northwest corner of Structure 2. Soot blackened the preserved portion of the northwest corner, and this constitutes the primary evidence for the existence of this feature. Ash was also present, but had been scattered by rodents. A stone in the north wall was broken open, and it was found to be oxidized beneath the soot-blackened surface, indicating that a fire had been built in Feature 1, probably for heat, light, and (possibly) cooking. A sample of the ash was collected for flotation analysis (Chapter 16, this volume).

        Roof. No evidence was found of the Structure 2 roof or roof-support features. The thickness of the walls suggests that the walls themselves supported the roof. The absence of a roof fall stratum above the floor suggests that the roof timbers may have been salvaged when the room was abandoned.

Dating

No absolute dates were obtained for Structure 2. Structure 2 is aligned with the masonry pit structure, Structure 1, and the masonry tower, Structure 4. Together these represent the main architectural features that date to the Pueblo III occupation of Lillian's Site. Tree-ring and archaeomagnetic dates from Structure 1 indicate that this Pueblo III occupation dates to sometime between A.D. 1175 and 1250, with the cutting dates between A.D. 1207 and 1211 and the latest tree-ring date at A.D. 1214 vv. The pottery found in the fill of Structure 2 is consistent with these dates; decorated white wares that could be identified to traditional type include Mesa Verde Black-on-white and McElmo Black-on-white (Mesa Verde Black-on-white pottery was not made until after A.D. 1180 [Wilson 1991]). The remaining decorated white wares, which could not be assigned a specific traditional type, were all decorated with carbon paint, indicating that they, too, date to the Pueblo III period.

Stratigraphy

Structure 2 did not burn. Wall fall in a silt loam matrix filled the entire structure, beginning at the same elevation as the base of the Structure 2 masonry walls (Figure 3.8). Tools found in the wall fall stratum include two modified flakes, a core, a modified sherd, a piece of ground mineral identified as gypsum/calcite/barite, and a side-notched, Pueblo III-style projectile point.

The last 10 cm of fill was provenienced as Stratum 2 during excavation because wall fall decreased in frequency and because pockets of ash were encountered. When the Structure 2 profile was drawn, no stratigraphic break could be seen between this last 10 cm and the upper fill. Structure 2 fill stopped at the base of the Structure 2 walls. A peckingstone and a modified sherd were the only tools in this last 10 cm of fill.

Structure 3 fill lay directly below the Structure 2 walls and wall fall. The upper portions of the Structure 2 walls appear to have collapsed directly onto the Structure 2 floor, which indicates that the collapse occurred shortly after the structure was abandoned. The absence of any roof fall in Structure 2 suggests that the roof timbers may have been salvaged. The salvaging of roof timbers may have partly destroyed the upper walls.

There is little evidence of postabandonment use of Structure 2. No surfaces were discovered in the wall fall, artifacts were scarce, and no features were recorded. The tools in Structure 2 fill may have been refuse thrown into the abandoned structure as it was collapsing.

Structure 2 Interpretations

Structure 2 is a masonry room constructed above an earlier post-and-adobe room. Structure 2 dates to the Pueblo III occupation of Lillian's Site. Pottery in the fill provides the only means of dating this room; the presence of Mesa Verde Black-on-white indicates that the structure filled after A.D. 1180.

A burned spot was found in the corner of Structure 2. Although heavily disturbed, this feature does not appear to be a pit. Instead, a fire appears to have been built on the Structure 2 floor. Thermal features are often associated with rooms interpreted as habitation areas or living rooms (Adams 1983b; Dean 1969, 1970; Hill 1970; Rohn 1965, 1971).

Stratigraphy indicates that Structure 2 collapsed shortly after abandonment. Collapse of the upper walls may have resulted from the salvaging of roof timbers, and not simply natural erosion. This interpretation, however, is based on limited testing.

Structure 3

Structure 3 is a post-and-adobe room discovered beneath Structure 2 (Figure 3.7 and Figure 3.8). Burned adobe and burned wood in the fill of Structure 3 indicate that the room burned. Structure 3 fill appears to have been leveled to create the floor of Structure 2.

Structure 3 walls were difficult to identify. This was partly because only a limited area was exposed during testing and partly because Structure 2 lay on the top of Structure 3. Structure 2 construction may have destroyed portions of the Structure 3 walls as well. It was not possible to determine the dimensions of Structure 3 on the basis of limited testing.

Construction

        Walls. Identification of Structure 3 as a post-and-adobe room is an inference made on the basis of indirect evidence. An upright slab appeared to be part of the west wall of Structure 3; upright slabs are usually thought to be associated with walls that are primarily post-and-adobe construction and not full masonry walls (Brew 1946:215-226). Fill above the floor of Structure 3 contains charcoal, burned beams, and adobe with beam impressions, but little sandstone. The adobe and burned beams may have been entirely from the Structure 3 roof, but they may have been from the walls as well.

        Floor. The floor of Structure 3 is an abrupt stratigraphic contact between the burned superstructure collapse and sterile B-horizon sediments (Figure 3.8). The floor was constructed by leveling off these sterile sediments, and subsequent use produced a compacted surface. Burned superstructure collapse covered the floor, oxidizing and further hardening the floor.

Two isolated flakes and a cluster of 16 flakes lay directly on the floor. These flakes appeared to be of the same raw material and together may represent a stone-tool reduction area within Structure 3.

        Roof. A number of beams lying perpendicular to one another were found just above the floor, and these probably are the remains of the fallen roof. Smaller beams, about 2 cm in diameter and apparently shaped, lay beneath slightly larger beams.

A beam partly encased in adobe was found 40 cm above the floor. This beam was a minimum of 6 cm in diameter, and it could be from either the roof or the wall.

Dating

Twenty-two tree-ring samples were submitted for analysis, but only four specimens, all juniper, produced dates. These dates are 499 vv, 543 ++vv, 560 vv, and 1075 vv. Mancos Black-on-white pottery was found in Structure 3 fill, indicating that the latest of these dates most accurately dates the structure. Having dates separated by over 500 years in the same deposit is a problem, but this is best explained by the fact that they are not cutting dates.

Post-and-adobe rooms in the Mesa Verde region often date to the Basketmaker III and Pueblo I periods. It is possible that Structure 3 dates to one of these earlier periods and that the wood that yielded the A.D. 1075 vv date was somehow mixed into the earlier burned roof fall during the later occupations of Lillian's Site. Post-and-adobe architecture, however, is not uncommon in Pueblo II contexts. Numerous post-and-adobe roomblocks dating between A.D. 1000 and 1100 have been documented 20 km (12 mi) north of Lillian's Site (Kuckelman and Morris 1988).

Test excavations revealed unburned posts beneath the masonry walls elsewhere at Lillian's Site (Features 1, 2, and 4 of Sampling Stratum 1). These posts, located over a distance of 10 m, indicate that Structure 3 may have been part of a post-and-adobe roomblock, and not an isolated room. Decorated pottery in the Structure 3 fill includes one Late White Painted sherd with mineral paint and one Late White Painted sherd with carbon paint. In addition, there were three Late White Unpainted sherds and seven corrugated sherds. Possible earlier sherds include four Indeterminate Plain Gray sherds. Thus, Structure 3 could date to either the Basketmaker III or Pueblo I periods, but the tree-ring, pottery, and architectural evidence suggests that Structure 3 was a part of a Pueblo II occupation that dates to the late A.D. 1000s.

Stratigraphy

A thin lens of ash covered portions of the Structure 3 floor, along with small pieces of charcoal, larger beams, and burned adobe. The fire that destroyed Structure 3 appears to have occurred at the time of abandonment.

The burned structural collapse lay in a matrix of unburned sediments that varied in color from light brown to red brown, and in texture from sandy loam to clay loam. Postabandonment mixing of the burned structural collapse with wind-deposited sediments may have produced the observed variation in color and texture.

The stratum containing the burned structural collapse was 41 cm thick. The floor and walls of Structure 2 lay on top of this stratum. Therefore Structure 3 fill may have been leveled to build Structure 2 (Figure 3.8).

Structure 3 Interpretations

Structure 3 is interpreted as a post-and-adobe room located beneath a masonry room, Structure 2. Unburned posts were found beneath the masonry walls elsewhere in Sampling Stratum 1, which suggests that Structure 3 was a part of a post-and-adobe roomblock, and not an isolated structure. Tree-ring dates indicate that Structure 3 was constructed after A.D. 1075. Late White Painted pottery with carbon paint designs was found in the fill of Structure 3, supporting the interpretation that this room is a part of a Pueblo II occupation of Lillian's Site. This Pueblo II architecture at Lillian's Site is similar to the post-and-adobe remains found at G and G Hamlet and Kenzie Dawn Hamlet (Chapters 2 and 5, this volume). Excavations 20 km (12 mi) north of the Sand Canyon locality indicate that post-and-adobe architecture may be common in much of the Montezuma Valley during the A.D. 1000s (Kuckelman and Morris 1988).

The West Roomblock Unit

Probability sampling unit 102N/95E overlapped the northeast corner of the western roomblock unit and the curved wall that abutted both the western and eastern roomblock units. Judgmentally located 1-×-1-m units were excavated north and south of 102N/95E to adequately interpret these walls and the associated stratigraphy (Figure 3.5 and Figure 3.7).

Construction

        Walls. Excavation in 102N/95E exposed the top of the northeast corner of the west roomblock unit and the exterior face of the east wall. This semicoursed wall is constructed with sandstone blocks; it stands three courses (36 cm) high and measures 32 cm wide. Mortar joints are 2 cm thick, and there is no chinking. The wall is double-stone wide in cross section, and, like the Structure 2 walls, small stones are used on the interior face and large stones on the exterior. Corner stones are extensively shaped with pecking, whereas minimal pecking and flaking was observed on the other stones. The northeast corner of the west roomblock unit appears to be bonded.

Stratigraphy

The west roomblock unit is built on a stratum of fill 28 cm thick. This fill contained unburned adobe and unburned wood and appears to be collapse from a post-and-adobe room. A post hole, Feature 4, was found in 102N/95E at the contact between the fill and the sterile sediments. This post hole may be a part of a wall of a post-and-adobe room.

West Roomblock Unit Interpretations

The west roomblock unit is interpreted as part of the Pueblo III occupation of Lillian's Site. The chronological relationship between the east and west roomblock units is not clear. A curved wall connects the east and west roomblock units. This wall abuts both the east and west roomblocks, indicating that its construction postdates the construction of the two roomblocks. It seems unlikely that the curved wall would have connected the two roomblock units if one were abandoned; therefore, at least partial contemporaneity of the two units is suggested.

Wall Connecting the East and West Roomblock Units

Excavation in 102N/95E also uncovered a curved wall running between the east and west roomblock units. Observations made during wall trenching and excavation indicate that the wall does not enclose a room. Further excavation would be required to confirm this interpretation.

Construction

        Wall. The semicoursed sandstone block wall is two to three courses high (30 cm) and 40 cm wide. Mortar joints are 2 cm thick, and chinking is absent. The wall is double-stone wide in cross section, with small stones on the south side of the wall and large stones on the north side. Elsewhere on the site, small stones are on the interior walls of structures, but a structure does not appear to exist to the south of this wall.

The wall stones show no evidence of shaping. Like the other masonry walls in Sampling Stratum 1, this wall was constructed on a stratum of fill from an earlier occupation.

Dating

The curved wall abuts the northeast corner of the west roomblock unit, and it also appears to abut the southwest corner of the east roomblock unit. Therefore, the curved wall was built sometime after the construction of both roomblock units. Elevations at the wall bases provide a clue as to when the walls were constructed relative to one another. The base of the curved wall was 99.74 m, the base of the west roomblock unit wall was 99.72 m, and the base of the Structure 2 wall in the east roomblock unit was 99.67 m. Presumably, the longer the time interval between construction events, the greater the change in the prehistoric ground surface (the accumulation of cultural fill during the occupation of the site would result in later constructions being built on a higher ground surface). The similarity in elevation between the bases of the walls described above may indicate that little time separated the construction of the two roomblocks.

Stratigraphy

Stratigraphy along the 95E line from 101N to 104N is illustrated in Figure 3.9. Deposits north of the curved wall consist of three strata. Stratum 1 (34 cm) is the dense wall fall in a brown silt loam matrix; this deposit is clearly the fallen walls of the Pueblo III roomblock.

Stratum 2 (36 cm) contains less wall fall than Stratum 1 and is a yellow brown to dark yellow brown silty clay loam containing abundant charcoal and tiny (1 to 5 mm) white flecks of calcium carbonate. The boundary between Strata 1 and 2 is clear to gradual. Stratum 2 probably accumulated both during and after the use of the Pueblo III roomblock.

Stratum 3 is a silty clay loam that differs from Stratum 2 in that it has fewer inclusions. The boundary between Strata 2 and 3, marked by a distinct lens of clay, is abrupt and smooth. The clay lens may be a surface; if so, it may have been the surface in use at the same time as the masonry roomblocks (the lens lies 8 cm below the base of the masonry). This possible surface is cut by either a feature or a large disturbance, making the interpretation of the clay lens and its relationship to the masonry wall unclear. The base of Stratum 3 is the contact with the sterile B horizon.

The loose fill adjacent to the north side of the curved wall is difficult to interpret--it could be either a cultural feature or a large animal burrow. The appearance of recent mixing and the fact that one area within the loose fill clearly was an animal burrow favors the latter interpretation; however, the extent of the loose, mixed fill was larger than that of the burrow, and there was not a clear, abrupt boundary between it and Strata 2 and 3. Fragmented human bones (parts of a cranium and the end of a long bone) and five large corrugated sherds that fit together were found at the base of this disturbance. The human bones do not represent an articulated burial, although more bones could have been located outside of the excavation unit.

Two strata were identified south of the masonry wall (Figure 3.9). Stratum 1 is dense wall fall in a silt loam matrix. The boundary between Strata 1 and 2 is at the same elevation as the base of the masonry walls and is interpreted as the courtyard (Nonstructure 2) surface in use during the Pueblo III occupation of the site. One feature, described below, was found on this surface.

Stratum 2 is a mixed deposit consisting primarily of brown silt loam. Masonry wall fall is absent, but unburned adobe, unburned wood, and white flecks of calcium carbonate are present. Stratum 2 is interpreted as structural collapse of a post-and-adobe roomblock, part of an earlier occupation of Lillian's Site. The base of Stratum 2 is the sterile B horizon.

        Feature 3 (Burned Spot). A burned spot, Feature 3 of Sampling Stratum 1, was recorded on the courtyard surface. Juniper charcoal was the only botanical material recovered from a flotation sample collected from Feature 3 (Chapter 16, this volume). Feature 3 appears to be the remains of a small fire built on the Pueblo III courtyard surface.

Structure 4

Structure 4 is the masonry tower; it is discussed here because the tower falls within Sampling Stratum 1. No Sampling Stratum 1 sampling units were excavated inside the tower, but two 1-×-1-m probability squares (87N/91E and 87N/96E) in Sampling Stratum 4 fell just outside the tower. These sampling units exposed the exterior tower wall at the southwest and southeast corners (Figure 3.4 and Figure 3.5).

Construction

Structure 4 dimensions can be approximated using the exposed corners. The outside diameter of Structure 4 is approximately 5.00 m. Structure 4 wall thickness was not observed, but tower wall thickness at Shorlene's, Roy's, and Troy's sites was 41, 47, and 69 cm respectively. If Structure 4 wall thickness is estimated at 50 cm, the interior tower diameter would measure 4.00 m.

        Walls. Structure 4 walls are a maximum of 90 cm high (at the southwest corner) and a minimum of 60 cm (at the southeast corner). Rubble surrounding the tower is extensive (wall fall from 87N/91E measures .94 m³, and wall fall from 87N/96E measures .65 m³). The interior of the tower is assumed to be filled with rubble as well. An especially dense trail of rubble comes off the northeast corner of the tower. This pile of rubble differs from the remainder of the tower rubble in that it is not overgrown with sagebrush and the rocks are not heavily covered with lichen. This portion of the wall may have been toppled during the chaining operation in the 1960s. The volume of wall fall suggests that the tower once stood more than one story high.

To estimate the original height of the tower, we added the volume of the existing wall and the total volume of the wall fall. The following estimates, based on field observations, are used in the calculations:

1. Wall thickness is 50 cm.
2. Radius to the exterior of the tower wall is 2.5 m.
3. Radius to the interior of the wall is 2 m.
4. Average existing wall height is 75 cm.
5. The tower and associated wall fall cover 36 m².
6. Wall fall volume per m² surface area is .745 m³.

Using these figures, we calculate that the surface area of the tower walls is 7.07 m². This is subtracted from the 36 m² to get the area where wall fall was present. This figure, 28.93 m², can be multiplied by .745 to produce the total volume of wall fall; the result is 21.55 m³. Volume of the existing tower walls is 5.30 m. Adding the volumes together, and converting volume to wall height results in an estimate of 3.89 m for the original tower walls.

The purpose of this reconstruction was to determine if the tower was more than one story. In all cases, the estimates used to reconstruct wall height are conservative. Therefore, this reconstruction indicates that the tower at Lillian's Site was probably two stories high.

Sampling Stratum 1 Summary and Interpretations

Sampling Stratum 1 is the sampling stratum that includes the masonry surface architecture. Excavations in the masonry roomblock include two randomly located sampling units and four judgmentally located sampling units. These excavations documented separate east and west roomblock units connected by a curved wall. The exterior of the masonry tower was also exposed. The masonry roomblocks and tower are associated with a Pueblo III occupation of the site. Artifacts associated with this occupation include Mesa Verde Black-on-white pottery, indicating that the occupation postdates A.D. 1180 (Wilson 1991). One masonry room, Structure 2, was tested. Structure 2 had a burned spot in the corner of the room and may have functioned as a living room. The tower, Structure 4, appears to have been two stories high.

Partial contemporaneity of the east and west roomblock units is suggested by wall abutments and elevations of the bases of the masonry walls. The east roomblock unit may have been constructed first, as it is aligned with the pit structure, the tower, and the midden at Lillian's Site.

Standing walls in the east roomblock unit were higher (58 cm) than those in the west roomblock unit (36 cm). Wall fall volumes in the three 1-×-1-m test pits in each roomblock were averaged to get an estimate of wall fall/m². Wall fall volume in the 1-×-3-m excavations in the east roomblock unit was greater (.80 m³) than wall fall volume in the 1-×-3-m excavations in the west roomblock unit (.50 m³). As noted above, an argument can be made that the east roomblock was constructed first, and therefore the lower walls and less dense wall fall in the west roomblock cannot be attributed to salvaging stone to build the east roomblock.

Wall height for the east roomblock unit was reconstructed in the same way as the Structure 2 wall height was reconstructed. Total wall fall in the east roomblock was estimated by subtracting the area of the standing walls from the size of the east roomblock rubble mound, then multiplying this area by the estimate of wall fall/m². The total wall fall was added to the height of the standing walls. This produced a reconstructed wall height of 1.28 m for the east roomblock walls. The west roomblock unit had less wall fall and lower standing walls, so the reconstructed wall height for the west roomblock would be lower than 1.28 m.

This wall height reconstruction indicates that there is not enough rubble present to account for full height, single-story masonry walls. Several explanations could account for this. First, not all of the rubble associated with the masonry roomblock may be visible on the surface (for example, some of the rubble may have fallen into the pit structure fill). Testing in the pit structure, however, did not encounter massive deposits of wall fall. Second, the rooms in the roomblock units may not have been full-height masonry walls; upper walls may have been constructed with adobe. Finally, stones from the masonry roomblocks may have been salvaged for construction elsewhere after Lillian's Site was abandoned. It seems unlikely that upper walls in the roomblock units were constructed with anything other than masonry, as both roomblock units have walls that are double-stone wide. Thus, the most plausible explanation may be that roof timbers and stone from the roomblocks were salvaged for use at another site after Lillian's was abandoned.

A stratum of fill with cultural material was found beneath the base of the masonry roomblock in each of the excavation units. This deposit includes unburned posts and the structural collapse from a burned post-and-adobe room. On the basis of these observations, a post-and-adobe roomblock is believed to lay beneath the masonry roomblock. Structure 3 appears to be a burned room in this roomblock.

Dating the early post-and-adobe construction is difficult. Three of the four tree-ring dates from the burned post-and-adobe room fall between A.D. 499 and 560, but all are noncutting dates, which are difficult to interpret. The fourth tree-ring sample dates to A.D. 1075, and it is also a noncutting date.

Dating with pottery is similarly ambiguous, with a range of early types present. Pottery that predates the Pueblo III period, found in levels below the masonry roomblocks, include Early White Painted (N = 1), Late White Painted with mineral paint (N = 3), Mancos Black-on-white (N = 4), and Local Plain Gray (N = 13). In addition, Local Corrugated sherds (N = 40), which could date to either the Pueblo II or III periods, are present in these deposits. These sherds are predominantly Pueblo II pottery, but some pottery dating to the Basketmaker III or Pueblo I periods is also present (Early White Painted and Local Plain Gray). It should be noted that there was also Pueblo III pottery in these deposits below the roomblock (Pueblo III White Painted [N = 1] and Late White Painted with carbon paint [N = 6]). The deposits beneath the roomblock appear to be mixed, possibly due to leveling this stratum for use as the foundation for the masonry roomblock. The combination of the A.D. 1075 vv tree-ring date and the Mancos Black-on-white pottery indicates that the post-and-adobe roomblock most likely dates to the late Pueblo II period. Recent excavations 20 km (12 mi) north of the Sand Canyon locality uncovered several well-dated sites with post-and-adobe roomblocks and associated pit structures occupied between A.D. 1000 and 1100 (Kuckelman and Morris 1988).

Twelve stone tools were recovered from Sampling Stratum 1: one mano, one peckingstone, one projectile point, five cores, three modified flakes, and one other modified stone. One hundred and fifty-four pieces of chipped-stone debris were also recovered.

Pit Structure and Courtyard Areas (Sampling Strata 2 and 3)

Sampling Stratum 2 was defined as the pit structure depression, and it contains 26 possible sampling units. Three randomly selected sampling units (95N/99E, 97N/99E, and 99N/98E) and one judgmentally selected sampling unit (96N/99E) were excavated in Sampling Stratum 2. Each of these sampling units landed inside the pit structure and were therefore provenienced as Structure 1.

Sampling Stratum 3 was defined to include the courtyard area (provenienced in the field as Nonstructure 2). The courtyard surrounds Structure 1 and is bounded by the surface roomblocks and the tower (Figure 3.5). A total of 28 sampling units are present in this sampling stratum; four sampling units (96N/96E, 97N/96E, 98N/102E, and 96N/102E) were excavated.

Figure 3.10 illustrates Structure 1, the sampling units excavated within Structure 1, and the sampling units excavated outside Structure 1 in Sampling Stratum 3. The units excavated outside of Structure 1 are important, because they reveal significant differences between Structure 1 and the kivas excavated at Sand Canyon Pueblo (Bradley 1986, 1987, 1988a). Most of the excavated kivas at Sand Canyon Pueblo are above-ground structures that were built up from the bedrock surface; many were constructed within square enclosing structures and had associated corner rooms. What we found in the excavation units outside Structure 1 at Lillian's indicates that enclosing structures and corner rooms are absent at this site.

The Sampling Stratum 3 units landed just outside the pit structure, to the east and west. Excavation in each unit involved the removal of a single stratum of cultural fill, 20 to 30 cm thick, down to the stratigraphic contact with sterile sediment. This contact with sterile sediment is the level at which features could be defined and probably closely approximates the elevation of the courtyard surface.

Portions of three shallow pit features are present in sampling units 96N/102E, 98N/102E, and 96N/96E. Features 2 and 3 are located on the side of the square closest to Structure 1 and extend out of the sampling unit in the direction of the pit structure, sloping gradually down toward Structure 1. These features are probably areas where the plaza surface eroded into Structure 1 after the pit structure was abandoned and partly filled.

Structure 1

Structure 1 (Figure 3.11) is a masonry-lined pit structure that dates to the Pueblo III occupation of Lillian's Site. Three contiguous sampling units were excavated, uncovering the southern recess and the portion of the lower main chamber containing the deflector and the hearth. A fourth sampling unit uncovered the northwest corner of the pit structure.

Structure 1 dimensions can be inferred from the limited testing. The length of the pit structure at the level of the bench surface is approximately 5.50 m; the width is 4.80 m. The lower main chamber appears to be circular, with a diameter of 3.80 m. The main chamber floor is 1.92 cm below the estimated prehistoric ground surface.

Construction

        Walls. Structure 1 appears to be entirely lined with sandstone masonry. The lower main chamber lining wall measures 19 cm thick and 82 cm high (from the floor to the bench surface). This is a single-stone-wide, fully coursed masonry wall built on sterile sediments. Mortar joints are .5 to 2 cm thick, and chinking is absent. Mortar is a brown to red brown silty clay loam with no inclusions; the mortar is similar in color and texture to the sterile B-horizon sediments. Stone faces are minimally shaped with pecking and flaking. As many as four coats of plaster cover portions of the lower main chamber lining wall. The upper main chamber walls in the southern recess (Bench 6, Feature 1) also consist of horizontally coursed masonry. The upper main chamber walls in the northwest corner of Structure 1 (Bench 4, Feature 6) were constructed using a combination of horizontal coursing and an upright vertical slab. Walls above the bench are heavily blackened with soot.

        Feature 1 (Bench 6): Feature 1 is Bench 6, the southern recess (Figure 3.12). The east side of Bench 6 is defined by Feature 2, Pilaster 6. Feature 8 is a pit feature found on Bench 6. The Bench 6 surfaces will be discussed first, followed by discussions of Feature 8 and Feature 2.

Only a portion of Bench 6 was exposed during excavation. Full dimensions are unknown, but the distance from the front to the back of Bench 6 is over 1.24 m.

Two, and possibly three, surfaces were associated with Bench 6. Surface 1, the upper surface, was constructed with adobe plaster. Surface 1 did not appear to be burned. A flake was the only artifact associated with this surface.

Surface 2, lying 2.5 cm below Surface 1, was also constructed with adobe plaster. Surface 2 was flush with the base of the Bench 6 masonry, but 2 cm below the masonry of the lower main chamber lining wall. Surface 2 appeared to have faint traces of burning; therefore, Structure 1 may have first burned and subsequently may have been remodeled and reused. Beneath Surface 2, a large pit feature, Feature 8, was found.

        Feature 8 (Pit): Feature 8 is a pit feature located on Bench 6 that was capped by both surfaces. Feature 8 partly undercuts the east wall of Bench 6. Only a portion of Feature 8 was exposed in the excavation unit; this portion measures 52 cm wide and 33 cm deep. The walls of Feature 8 consist of sterile sediment (caliche). The fill is a single stratum of red brown silty clay loam, mottled with white flecks (calcium carbonate). A lens of this same fill, 2 cm thick, lay beneath Surface 2 and above the sterile caliche surface outside Feature 8. The caliche surface may represent a third surface associated with Bench 6. The filling of Feature 8 may relate to the Surface 2 construction of Bench 6.

Three clusters of small pieces of burned wood were found in the fill of Feature 8. Nine tree-ring samples were collected from these clusters. In one cluster, five samples may be from the same broken beam; two separate small beams were present in the other two clusters. Three of the tree-ring samples produced dates: A.D. 1058 vv, 1196 vv, and 1207 +r. These dates indicate that Feature 8 was part of Structure 1 and not an earlier structure or feature truncated by Structure 1.

During excavation, the base of Feature 8 displayed characteristics of a surface: overlying fill "popped" off, and the base had a gray-brown color as if it were use-compacted. No evidence of burning was observed on the walls or base of the feature.

Interpretation of Feature 8 is difficult. Tree-ring dates indicate that the feature was used as part of Structure 1 and was not constructed during an earlier occupation of Lillian's Site. Clearly, it was not in use when Structure 1 was abandoned, as the pit had been capped by two bench surfaces. The well-defined walls and base of Feature 8 indicate substantial use of this pit at some time. Feature 8 may have been deliberately filled during the construction of Surface 2 of Bench 6. The burned wood in the fill probably represents trash, as the walls and base of the feature were not burned. This burned wood may be associated with a fire in Structure 1 that took place early in the use of this structure. Additional evidence that supports the interpretation of an earlier fire in Structure 1 is presented below.

        Feature 2 (Pilaster 6): Feature 2, Pilaster 6, is a masonry pilaster. Pilaster 6 is partly collapsed, and only a portion of the feature is present within the sampling unit, making interpretation of the feature difficult. The pilaster consists of two parts: a masonry column that rises from the Structure 1 floor and the wall that defines the east wall of Bench 6. The latter is constructed on top of Surface 1 of Bench 6, and the masonry of the Bench 6 portion of this pilaster abuts the masonry column. It is possible that the two portions of this pilaster reflect remodeling of Structure 1.

The portion of the pilaster above Bench 6, Surface 1, is a maximum of nine courses high (83 cm) and a minimum of two courses high (18 cm). Wall fall from this construction rested directly on the bench surface, indicating that wall collapse occurred shortly after Structure 1 abandonment. The pilaster that defines the east wall of Bench 6 was heavily sooted. Plaster covered portions of the wall, and both the plaster and the stone behind the plaster were soot-blackened.

The portion of the pilaster that forms the east wall of Bench 6 abuts a masonry column built up from the floor of the main chamber. This same type of pilaster construction occurs at G and G Hamlet. Both the lower main chamber lining wall and the east wall of Bench 6 abut the pilaster, indicating that the masonry column was part of the initial construction of Structure 1. The columnar portion of Pilaster 6 is 17 courses high (1.35 m). Three coats of plaster covered the lower portion of Pilaster 6. Faint traces of the soot that covered the stones in the upper portion of the pilaster indicate that plaster did not completely cover Pilaster 6 during the entire use of Structure 1.

Examination of the interior of a broken stone from the columnar portion of Pilaster 6 revealed that it was oxidized 1 cm in from the soot-blackened face. Use of Structure 1 could have produced heavy soot-blackening, but use alone could not produce oxidation of the masonry walls. A fire must have occurred in Structure 1 as well.

There may have been a fire during the early use of Structure 1. Feature 8 apparently dates to this early use. The structure, including Bench 6, apparently was remodeled after this fire. This remodeling included the filling of Feature 8 and the construction of the portion of Pilaster 6 that forms the east edge of Bench 6.

        Feature 6 (Bench): Feature 6 is the bench in the northwest corner of the pit structure. Limited testing in Structure 1 makes it impossible to be certain, but this segment is probably Bench 4. The portion of Feature 6 uncovered in sampling unit 99N/98E measured 37 cm wide and 1.28 m long.

The bench surface was constructed with adobe plaster. Two flakes were found embedded in the bench surface. Ash, 1 to 2 cm thick, covered part of the bench surface. The upper lining wall of the bench was heavily soot-blackened. The combination of ash and soot-blackening is evidence for a fire on the bench surface before the abandonment of Structure 1. A piece of charcoal from this ash concentration was collected for tree-ring dating and produced a date of 1209 vv.

        Feature 7 (Pilaster): Feature 7 is a small corner of a pilaster, probably Pilaster 3. This pilaster forms one boundary of Feature 6 (bench). Feature 7 was constructed on top of the bench surface, using horizontally coursed masonry. Only two courses of masonry were observed. The pilaster may be collapsed, or the top of the pilaster may be intact outside the excavation unit.

Two coats of plaster covered the exposed portion of Feature 7. The inner coat was soot-blackened, the outer coat was not. This contradicts the evidence from Feature 6. The evidence for burning in Structure 1 is complicated; it is summarized below in the section on dating.

        Floor. The floor of Structure 1 is constructed of adobe plaster. Floor artifacts included a black (igneous) polishing/hammerstone that was ground, polished, flaked, and battered and a red polishing stone that was ground and polished. Both were found south of the deflector. A small animal bone lay on the floor in the northwest corner of Structure 1; unfortunately, this bone was lost before it could be analyzed.

A lens of ash was found 5 cm above the floor in the northwest corner of the structure. Three sandstone slabs covered the floor near the hearth, and pollen samples were collected beneath these slabs (see Chapter 17 for detailed reporting of these samples). Economic pollen found in the floor samples includes maize, cheno-am, beeweed, and purslane.

There may be a use-compacted surface on top of sterile sediments below the adobe surface in Structure 1, as evidenced by a thin gray lens seen in the profile. Further excavation in Structure 1 would be needed to determine if there is an earlier surface in Structure 1.

Two features, a hearth and a deflector, are associated with Surface 1 in Structure 1.

        Feature 5 (Hearth): Feature 5 is the Structure 1 hearth (Figure 3.13). This hearth was constructed by excavating a pit into the floor, and then lining the pit with adobe. The hearth measures 80 cm in diameter and 29 cm deep. A small portion of the hearth extends north, outside of the sampling unit.

Burning discolored the rim and upper walls of the hearth. Hearth fill consisted of an upper stratum (8 cm) of the same deposit that covered the structure floor. A lower stratum (21 cm) of ash filled the remainder of the hearth.

The ash was sampled with five 1-liter flotation samples, and the ash that remained was fine-screened. Three flotation samples were analyzed. Charcoal interpreted as fuel includes sagebrush juniper, and bitterbrush/cliffrose wood. Mormon tea, serviceberry/squawapple wood, pine bark scales, and corn cupules may also be the remains of fuel. Corn and purslane and lemonadeberry seeds are possible food resources recovered from the hearth (Chapter 16, this volume). The ash that was not collected for flotation analysis was fine screened. A core, 96 pieces of chipped-stone debris, six sherds, and numerous small animal bone fragments were recovered from the fine-screen residue.

        Feature 3 (Deflector): A masonry deflector (Feature 3) is located 14 cm south of the hearth. Only a portion of the deflector was exposed during excavation. The deflector is one stone wide, is constructed of fully coursed masonry, and is 15 cm thick and 68 cm (seven courses) high.

Traces of plaster covered the lower portion of the deflector, and faint traces of soot stained the plaster. Soot-blackening stained the upper courses of the deflector that were not covered with plaster; however, the soot on the deflector was not as thick as the soot on the upper main chamber walls.

        Roof. The Structure 1 roof was supported by pilasters, two of which were exposed by excavation. Most Mesa Verde-region masonry-lined pit structures have a six-pilaster roof-support system, and this is probably the case for Structure 1 as well. Structure 1 roof timbers were not found and may have been salvaged for use at another site when Lillian's Site was abandoned. Evidence for salvaging timbers from Structure 1 is presented in the section on stratigraphy.

Dating

The most unambiguous dating for Structure 1 comes from an archaeomagnetic sample from the hearth. This sample produced a date of A.D. 1175-1250. This date, and the presence of Mesa Verde Black-on-white pottery associated with Lillian's Site, suggests that the abandonment of Structure 1 postdates A.D. 1180.

Ninety-four tree-ring samples were submitted from the fill and floor of Structure 1, and another nine were submitted from the fill of Feature 8, a pit capped by the southern recess bench surface. Fifty-nine of these samples produced dates, and the results of this analysis are presented in Table 3.1.

The dates range from 854 vv to 1214 vv. One cutting date, 1211 rB, is present, and the 1207 +r and the two 1212 v dates may also represent cutting dates. The interpretation of these dates is difficult because the origin of the wood is unclear. A single stratum of burned roof fall is not present. Instead, burned wood was found throughout Strata 2 and 3. The first tree-ring sample was collected at 98.82 m elevation, and the last lay on the floor at 97.35 m. In addition, samples were found on the bench surface (Feature 6) and in a pit (Feature 8) capped by the southern recess surface.

There is no clear pattern to the distribution of the dates relative to the location of the tree-ring samples in Stratum 2; early and late dates are associated with both upper fill and deeply buried surfaces. The fact that dates from the early A.D. 1200s were yielded by samples collected from upper fill, floor-associated contexts, and capped features suggests that Structure 1 filled relatively quickly and that the source of charcoal was a structure or structures dating to the Pueblo III occupation. The processes that filled Structure 1 are explored in greater detail in the section on stratigraphy.

The tree-ring samples are all fragmentary and small, with no sample exceeding 6 cm in diameter. The samples are larger than the charcoal bits found in hearths, however, and there was no ash in Stratum 2. Therefore, the Stratum 2 charcoal appears to be fragments of structural wood rather than fuel refuse from hearths.

The presence of adobe in Stratum 2 supports the interpretation that the charcoal is from abandoned structures. Charcoal from the capped pit, Feature 8, and charcoal associated with the floor probably represents wood associated with Structure 1. It is not clear if the remainder of the charcoal in Stratum 2 comes entirely from Structure 1, or if wood from other structures could be present as well.

Two observations indicate that Structure 1 burned at the time of abandonment: the ash on the surface of the bench and the ash above the floor in the northwest corner of the structure. It may be that the larger roof beams were salvaged and the smaller closing material was burned when Structure 1 was abandoned.

There is also evidence that Structure 1 burned and was subsequently remodeled and reused. This evidence includes the tree-ring samples from Feature 8, the deliberate filling of that feature, and the capping of that feature by Surfaces 1 and 2 of Bench 6. The construction of Pilaster 6 also appears to indicate remodeling. The earliest portion of Pilaster 6, the column that rises from the Structure 1 floor, is oxidized. The oxidation of these stones may have been the result of an early fire in Structure 1. In addition, there are places where unsooted wall plaster covers soot-blackened plaster.

On the basis of this evidence, a use history can be constructed for Structure 1. The cutting date from the burned wood in Feature 8 indicates that the initial construction of Structure 1 may have taken place around A.D. 1207. At some point, the structure burned and was remodeled. The A.D. 1209 vv date from the surface of the bench, Feature 6, is probably associated with the remodeling of Structure 1. The cluster of cutting dates around A.D. 1211-1212 and the latest date of A.D. 1214 indicate that this remodeling took place in the early A.D. 1200s.

The tree-ring data are supported by the archaeomagnetic date from the hearth. The latter indicates that Structure 1 was used and abandoned in the period between A.D. 1175 and 1250. In addition, there is Mesa Verde Black-on-white pottery in the fill of Structure 2, indicating that it filled sometime after A.D. 1180 (Wilson 1991).

Stratigraphy

Figure 3.14 is a view of the stratigraphy of the north face of grid unit 97N/99E. Figure 3.15 illustrates the north-south stratigraphic profile through Structure 1. Structure 1 fill includes only three strata. Strata 2 and 3 are identical in terms of color and inclusions; only a slight change in texture allows the two strata to be distinguished. The strata are so similar that they were excavated as a single stratum that was subdivided into levels.

Stratum 3 lies on top of the Structure 1 floor. Stratum 3 is deepest where it covers the deflector and is 80 cm thick. The sediment consists of a brown sandy clay loam. Abundant adobe (burned and unburned), charcoal, calcium carbonate flecks, and small (less than 10 cm), medium (10 to 20 cm), and large (greater than 20 cm) pieces of sandstone are the inclusions present in Stratum 3. Total sandstone from the Structure 1 excavations measures .45 m³.

Stratum 2 is the same as Stratum 3, except the texture is a slightly finer silty clay loam. This stratum is basin-shaped and a maximum of 1.2 m thick. The inclusions present in Stratum 3 were also present in Stratum 2. A rock concentration in Stratum 2 in 99N/98E had stones that appeared to be several courses of a wall that fell intact. This was provenienced as Feature 4 of Structure 1.

The Structure 1 profile was examined closely for evidence of microstratigraphy, especially the sorting by particle size that results from wind and water deposition. Wind and water deposits both "fine upward," resulting in coarse material settling toward the bottom and finer material rising toward the top for a single depositional event (for example, sediment washing into a structure after a rainstorm). This type of microstratigraphy is absent in the fill of Structure 1.

A thin (1 cm), discontinuous lens of burned sediment is the only lensing present in the northeast corner of 97N/99E. This lens is located 18 cm above the boundary between Strata 2 and 3. No oxidation is present on the top or bottom of this lens. It is not clear if it represents an in-place burn or burned material thrown in as refuse.

Strata 2 and 3 do not appear to be the result of wind or water deposition. Colluvial deposition or cultural deposition are the other possibilities. The favored interpretation is cultural deposition. This interpretation is based on modeling the difference between gradual filling by natural processes and rapid filling by cultural processes.

If Structure 1 was abandoned and allowed to fill naturally, the process would have been gradual, taking a minimum of several years (Kilby 1998). This gradual process would result in many distinct depositional events. The roof would gradually erode, wind would blow sediment into the structure, rains would wash sediment into the structure, and walls might slump. These separate depositional events would produce stratification within the fill sequence, which is not the case in Structure 1.

Strata 2 and 3 contain the same inclusions from top to bottom, and many of these inclusions are too large to have been deposited by wind or water. The inclusions and the matrix surrounding them are materials that were transported into the structure as dry aggregates. The inclusions have angular surfaces, indicating that they were not transported a great distance, were covered quickly, and were not subject to erosion. The tabular inclusions (for example, sandstone) are oriented at dissimilar angles; natural deposition would tend to orient these inclusions similarly and parallel to the bedding planes when they were deposited.

The absence of stratification, and the size, condition, and orientation of the inclusions implies rapid deposition. Rapid deposition of such a large stratum could be produced only by gravity or by human deposition. There is little opportunity for colluvial deposition, the only sources being the rooms above Structure 1 and the Structure 1 walls themselves. Colluvial deposition would also be a relatively gradual process, and therefore mixing with other types of depositional events would occur. Cultural deposition--that is, the deliberate fill of Structure 1 with Strata 2 and 3 at the time of Structure 1 abandonment--is believed to be the best explanation.

The presence of adobe, wood, and sandstone indicates that Strata 2 and 3 may be material from dismantled structures. Roof beams from Structure 1 may have been salvaged, and the sediment and closing material from the roof may have been deposited in Structure 1 as a result. Roof beams from the roomblock may have been salvaged also, and the detritus from these roofs deposited into Structure 1 as well.

The artifacts from Strata 2 and 3 in Structure 1 also indicate that these strata were culturally deposited. Complete and fragmentary tools found in this fill include two two-hand manos, one metate, one core, one peckingstone, one hammerstone, one polishing stone, one projectile point, one biface, five modified sherds, and two modified bones. Pottery from Strata 2 and 3 include 302 sherds. Decorated sherds from the fill include every white ware type recognized in the Crow Canyon analysis system. Early White Painted (1), Pueblo II White Painted (1), Late White Painted with both mineral and carbon paint (34), Pueblo III White Painted (12), Chapin Black-on-white (1), Piedra Black-on-white (1), Mancos Black-on-white (2), McElmo Black-on-white (1), and Mesa Verde Black-on-white (9) are all present in these strata. Gray wares are also represented by a wide range of types including Chapin Gray (2), Mancos Gray (1), and Mesa Verde Corrugated (1), with corrugated body sherds (130) being the most common type of sherd in these strata. The wide range of tools and pottery types is best accounted for by cultural deposition.

Stratum 1 fills the remainder of Structure 1. Stratum 1 varies in color from brown to dark grayish brown to black, and in texture from loam to silt loam. Virtually no inclusions were present in Stratum 1, except for a few small pieces of sandstone and artifacts. A total of 85 sherds was recovered, including Plain Gray (13), Corrugated Gray (39), Late White Painted (13) and Late White Unpainted (20). Stone tools include two modified flakes and one gizzard stone. Sediments are sorted into laminae in this stratum, such that darker lenses with greater concentrations of organic material alternate with lighter-colored lenses containing less organic material. On the basis of these observations, Stratum 1 is interpreted as a natural postabandonment deposit.

Structure 1 Interpretations

Structure 1 is a masonry-lined pit structure, the type of structure traditionally interpreted as a Mesa Verde-style kiva (e.g., Dutton 1938:61; Vivian 1965:22). A functional interpretation of Structure 1 is difficult, given the limited excavation.

The consistent form of Mesa Verde-style kivas raises the possibility that they had important symbolic-ritual aspects (Lipe 1989). The size of Lillian's Site, however, indicates that Structure 1 probably was not used in the same manner as a historic-period kiva, that is, as a structure relatively specialized for ceremonial activity and used by large groups of people not necessarily related by kinship. Structure 1 also lacks the complex kiva architecture documented for Sand Canyon Pueblo kivas, where kivas are enclosed in square structures and have associated corner rooms.

Structure 1 probably had a more generalized function, one that may have included both domestic and ritual activities. It is possible, for example, that the two roomblock units at Lillian's Site were used by distinct households or by separate infrahousehold groups (for example, groups from successive generations) that formed a single, large, multigenerational household. These groups probably shared the use of Structure 1. Shared use is a means by which Structure 1 may have functioned as an integrative facility, one in which both domestic and ritual activities took place.

Inner Periphery (Sampling Strata 4 and 6)

Sampling Strata 4 and 6 compose the inner periphery of the site. Sampling Stratum 4 is characterized by surface remains that include a low topographic rise and a light scatter of sandstone rubble. A total of 108 sampling units are present in this sampling stratum; four sampling units (87N/91E, 87N/96E, 91N/101E, and 87N/88E) were excavated.

Two sampling units (87N/91E and 87N/96E) have been discussed briefly in the description of the tower, Structure 4. The tower will not be discussed here. A wall abutting the tower was found in 87N/96E. This construction appears to be part of a low wall that encloses a circular area next to the tower. This is discussed below as Nonstructure 3. Another portion of Nonstructure 3 may be present in 91N/101E as well.

All Sampling Stratum 4 sampling units were excavated to sterile. No features, except the walls in 87N/96E and 91N/101E, were found in the Sampling Stratum 4 units.

Sampling Stratum 6 was defined on the basis of a low topographic rise and moderate artifact density. The masonry surface rubble in Sampling Stratum 4 was almost absent in Sampling Stratum 6. Two hundred twelve sampling units are present in Sampling Stratum 6; four sampling units (88N/102E, 105N/86E, 90N/103E, and 91N/107E) were excavated. Sampling Stratum 6 excavations may have also uncovered a possible wall alignment associated with Nonstructure 3 in 90N/103E.

The sampling units in Sampling Stratum 6 were excavated in natural strata down to sterile sediments. A single stratum, the modern A horizon, was present in each of the sampling units. This stratum was shallow, ranging from 13 to 30 cm.

A single feature was found in Sampling Stratum 6 excavations.

Feature 1 (Pit)

Feature 1 is a portion of a feature located at the contact with sterile sediments in 91N/107E. This feature is an amorphous pit, 56 cm in diameter and 16 cm deep. Brown silt loam with flecks of charcoal filled the feature. No artifacts were recovered, and the temporal affiliation of Feature 1 is uncertain.

Plaza (Nonstructure 3)

A low masonry wall was found abutting the tower, Structure 4. This wall appears to enclose a circular area interpreted as a small plaza (provenienced in the field as Nonstructure 3). This interpretation is based on a wall found in 87N/96E, possible walls in 91N/101E and 90N/103E, and surface alignments of rubble that extended out from the walls exposed by excavation. Limited testing makes the interpretation of this area as a partly enclosed plaza preliminary, and further excavation would be needed to clarify details of the construction and use of this area.

Construction

        Walls. The boundaries of this plaza are defined by two arcs of sandstone rubble that abut the southeast and northeast corners of the tower. These rubble alignments are most dense near the tower (where they are mixed with tower wall fall), and become gradually less dense away from the tower. The two rubble arcs enclose a circular area, but the arcs do not meet; a gap of approximately 4 m on the east side of the enclosure may be an opening into the plaza. The area within the arcs of rubble is devoid of rubble.

The enclosing rubble makes the area in the center look like a depression. This was recorded as a possible pit structure by the survey crew. Topographic mapping shows that the area in the center of the enclosure is not lower than the ground surface immediately outside the rubble arcs. The center of the enclosure was probed with an auger to test for the presence of a pit structure, and sterile sediments were found 20 cm below the modern ground surface.

The wall found in 87N/96E abuts the tower wall and runs in the direction of the rubble arc that defines the south side of plaza. This wall was three to four courses (27 cm) high and a single course (24 cm) wide.

An arc of rubble also defines the north side of the plaza. A masonry alignment within the arc of rubble was also found in 91N/101E, and this alignment continued to the east, where it was exposed again by excavations in 90N/103E. The alignment consists of stones that appear to be set end to end, but the stones are only one course high, so it is difficult to evaluate them as a wall. The alignment continues back toward the tower until it disappears under the rubble of the tower.

The area enclosed by the rubble arcs measures approximately 6.80 m in diameter. Features similar to this plaza have been identified on Pueblo III sites elsewhere in the Mesa Verde region and interpreted as dance plazas (Lange et al. 1986:28).

Inner Periphery Artifacts

Artifacts recovered from Sampling Strata 4 and 6 are summarized in tables that appear at the end of this chapter, in the material culture chapter, and in the faunal remains chapter. Sherd totals for Sampling Strata 4 and 6 were 192 and 178, respectively. There were 28 pieces of chipped-stone debris from Sampling Stratum 4 and 39 pieces from Sampling Stratum 6. The inventory of complete and fragmentary tools from these sampling strata includes one abrader, three cores, four modified flakes, one other chipped-stone tool, one other modified stone, and two modified sherds. In addition, 50 pieces of chipped-stone debris were recovered from the inner periphery excavations.

The inner periphery at Lillian's Site is on a relatively high point; the roomblock and tower rubble mounds are the only areas above the inner periphery and are therefore the only places from which materials could have eroded into the inner periphery. It seems unlikely that all of the inner periphery artifacts were deposited as a result of erosion from the roomblock and tower areas.

The assemblage of decorated white wares from the inner periphery is dominated by Pueblo III pottery: 45 Late White Painted sherds with organic paint, four Pueblo III White Painted sherds, and three Mesa Verde Black-on-white sherds date to the Pueblo III occupation of Lillian's Site; seven Early White Painted and eight Late White Painted sherds with mineral paint are the only decorated white wares that probably predate the Pueblo III occupation. This pattern contrasts with that of the roomblock, pit structure, and courtyard sampling strata, where greater numbers and varieties of pre-Pueblo III sherds were found.

Inner Periphery Summary and Interpretations

Excavations in Sampling Strata 4 and 6 revealed slightly greater artifact densities than in the courtyard and outer periphery sampling strata. However, the artifact density of the inner periphery at Lillian's Site was not as high as the artifact density of the inner peripheries at the other mesa-top sites. On the basis of artifact density and the types of decorated white wares present in the inner periphery at Lillian's, deposition in these sampling strata is interpreted as refuse associated with the Pueblo III use of the site.

Ethnoarchaeological studies of site-formation processes have described how "toft zones" develop around the edges of courtyards and patios (Arnold 1990:918; Deal 1985:262; Hayden and Cannon 1983:126). Refuse accumulates in these areas as a result of sweeping and other maintenance activities designed to keep courtyards free of refuse; these behaviors could perhaps account for the accumulation of refuse in the inner periphery. An interesting question for further study is whether the refuse in the inner periphery is somehow distinct from the refuse that gets deposited in the midden.

Midden (Sampling Stratum 5)

Sampling Stratum 5 is the midden sampling stratum. Because the midden is a culturally bounded, nonarchitectural study unit, it is provenienced as Nonstructure 1. The midden is a 60-cm-high mound with a high density of artifacts on the surface; the boundaries are clear (Figure 3.5). Several pot hunters' holes are present in the midden.

A total of 177 sampling units make up this sampling stratum; 10 were excavated: 76N/91E, 73N/87E, 77N/87E, 68N/90E, 69N/92E, 69N/85E, 70N/95E, 72N/95E, 72N/82E, and 78N/96E. The midden was heavily sampled to obtain enough artifacts and ecofacts for intersite comparisons.

Sampling units are located both in the center and on the edges of the midden. Excavation in each sampling unit stopped at the contact with sterile sediments. Sampling units are 57 cm deep in the center of the midden and 25 to 30 cm deep on the edge of the midden.

The midden was excavated in two strata. The upper stratum, Stratum 1, is a loosely compacted, structureless, brown silt loam. This stratum contains the majority of the artifacts and continues down to just above the sterile sediments. Stratum 1 ranges from 20 to 47 cm thick; where Stratum 1 is more than 20 cm thick, it is subdivided into levels. Stratum 1 sediments are darker brown than sediments outside the midden, but the ashy, charcoal-rich deposits that characterize many middens (for example, the midden at the Green Lizard site) are not present in Nonstructure 1 at Lillian's Site.

There is no evidence of stratification within Stratum 1. This is surprising, since Lillian's is a multiple-component site. We expected to see multiple deposits of secondary refuse separated by naturally deposited strata that accumulated during periods of abandonment.

Stratum 2 is a red brown silty clay loam containing few artifacts. This stratum lies between the secondary refuse above and the sterile sediments below. The prehistoric ground surface before the occupation of the site is located somewhere within Stratum 2.

Burial 1

A portion of a burial was found in the northwest corner of sampling unit 73N/87E. Because the majority of the burial was outside of the sampling unit, it was not excavated, but many observations concerning the burial were recorded.

The burial appears to be in a shallow pit. The human bone observed was not articulated, and the burial may have been disturbed by looters. Two long bones, the distal end of a left femur and the proximal end of the left tibia, lay 12 to 28 cm below the modern ground surface and extended 6 cm into the excavation unit. In addition to these long bones, which were left in place, a thoracic vertebra, a second phalanx, and two rib fragments were found. These smaller bones were removed during excavation so that the sampling unit could be completed, then were reburied adjacent to the long bones. The skeletal material is from an adult; the sex could not be determined.

Midden Artifacts

The majority of the artifacts recovered from Lillian's Site were obtained from excavation in the midden; 66.3 percent of all the pottery recovered from the site comes from the midden. Decorated white ware pottery that predates the Pueblo III occupation includes Piedra Black-on-white (N = 1), Early White Painted (N = 7), Mancos Black-on-white (N = 9), Pueblo II White Painted (N = 6), and Late White Painted with mineral paint (N = 58). Pueblo III decorated white ware pottery probably includes Late White Painted with carbon paint (N = 398), as well as the following types that are more confidently assigned to the Pueblo III period: Pueblo III White Painted (N = 124), McElmo Black-on-white (N = 8), and Mesa Verde Black-on-white (N = 25). Early gray wares include Moccasin Gray (N = 3), Indeterminate Neckbanded Gray (N = 1), Indeterminate Plain Gray (N = 80), and Mancos Corrugated (N = 9); late gray wares include Mesa Verde Corrugated (N = 6). Indeterminate Local Corrugated, which could be either Pueblo II or III, makes up the largest proportion of the gray ware sherds (N = 1,694). Although the majority of the sherds recovered are Pueblo III types, Pueblo I and Pueblo II pottery is present in substantial numbers as well. Early and late pottery is found in both Strata 1 and 2 of the midden, and where Stratum 1 was subdivided into levels, early and late pottery was present in each of the levels. This suggests that the midden has been heavily mixed by postabandonment processes such as chaining, looting, and rodent activity.

Complete and fragmentary tools found in the midden include one mano, seven abraders, one hammerstone, four peckingstones, 18 cores, two projectile points, two bifaces, three drills, 26 modified flakes, 43 modified sherds, one shaped sherd, 11 pieces of modified bone, one bone awl, and three bone awl fragments. A stone pendant fragment, a piece of shell, numerous gizzard stones, and pieces of eggshell were also recovered. Finally, 1,285 pieces of chipped-stone debris were recovered from the midden.

Flotation analysis was done on 13 samples from the midden (Chapter 16, this volume). Botanical remains interpreted as food and fuel were recovered. Seeds of cheno-am, Physalis longifolia, Opuntia, and Zea mays are interpreted as food resources. Juniper wood was the dominate charcoal, but pieces of Pinus, Purshia, Amelanchier/Peraphyllum, Quercus, and Artemisia charcoal were also found. Zea mays cupules are interpreted as evidence that corncobs were burned as fuel.

Midden Interpretations

The midden is a spatially discrete feature with well-defined boundaries; it is a trash mound, not sheet trash. The mound is aligned with the following major architectural units at Lillian's Site: the east roomblock unit, the pit structure (Structure 1), and the tower (Structure 4). This indicates that the trash mound is associated with the Pueblo III occupation at Lillian's Site. Artifacts from earlier occupations are mixed into the trash mound, indicating that trash from earlier occupations is present, but in small amounts.

A single burial was found during the midden testing. As is the case on many small Puebloan habitation sites, the midden functioned as a cemetery as well as the location for the disposal of refuse (Prudden 1903:13).

North Outer Periphery (Sampling Stratum 7)

Sampling Stratum 7 is the outer periphery sampling stratum at the north end of Lillian's Site. Surface remains associated with Sampling Stratum 7 include a light artifact scatter but no evidence of architectural features. Four hundred seventy-one sampling units are present in Sampling Stratum 7; four (104N/83E, 106N/106E, 113N/102E, and 83N/93E) were excavated.

All of the sampling units contain a single stratum, the modern A horizon. Stratum 1 was removed down to the sterile B horizon. This stratum was shallow, ranging from 12 to 20 cm.

Artifact density was light in all of the sampling units, except for 83N/93E, which was located 2 m north of the midden (Nonstructure 1). Here, sediments were darker and artifacts more abundant. For example, 136 of the 232 total sherds in Sampling Stratum 7 and 108 of the 132 pieces of chipped-stone debris were from 83N/93E. Three of the four tools recovered from Sampling Stratum 7 were found in this sampling unit as well, including a biface fragment, a projectile point fragment, and a metate fragment (a biface fragment was the only other tool found in Sampling Stratum 7). In contrast to the midden, the deposits in 83N/93E were relatively shallow; sterile sediment was encountered 20 cm below the modern ground surface.

No culturally bounded study units were identified in Sampling Stratum 7. Only one feature was found.

Feature 1 (Slab Cluster)

Feature 1 was discovered in the southwest corner of 106N/106E, at the contact with the sterile B horizon. Feature 1 consists of a group of sandstone slabs oriented horizontally--one large slab with eight smaller slabs around it. The slabs are carefully arranged, and they appear to have been set into the top of the sterile sediments. The slab cluster measures 60 × 60 cm and is 2 cm thick. No evidence of burning or wear was observed on the slabs. The slabs were removed to see if they capped a pit feature, but no pit was found. The function and temporal affiliation of Feature 1 are unclear.

South Outer Periphery (Sampling Stratum 8)

Sampling Stratum 8 is the outer periphery surrounding the midden at the south end of the site. Eight hundred eighty-six sampling units are present in Sampling Stratum 8; four (69N/82E, 70N/112E, 77N/113E, and 82N/114E) were excavated.

Excavation in each sampling unit involved the removal of Stratum 1 (brown silt) down to the contact with the sterile B horizon. Excavations were shallow, ranging from 12 to 17 cm in three of the sampling units; in 82N/114E, Stratum 1 was 35 cm deep. Artifacts were scarce: 84 sherds, 21 pieces of chipped-stone debris, four stone tools (one biface, one core, and two modified flakes), and two modified sherds make up the entire assemblage from Sampling Stratum 8. One feature was found in Sampling Stratum 8.

Feature 1 (Post Hole)

Feature 1, a post hole in 82N/114E, measures 19 cm long, 16 cm wide, and 22 cm deep. Rodents had disturbed the upper portion of the feature. Brown loam sediment, but no wood, filled the feature. Two pieces of charcoal were found just outside of Feature 1 at the contact between fill and sterile, and these yielded tree-ring dates of A.D. 899 vv and 964 vv. Burned adobe and charcoal were found in the fill above Feature 1. Therefore, this post hole may be part of a post-and-adobe architectural feature. Stockades surround small sites dating to the A.D. 1000s 20 km (12 mi) north of the Sand Canyon locality (Kuckelman and Morris 1988). Feature 1 may be part of a similar stockade.

Lillian's Site Artifacts

Artifact data are summarized in the material culture tables at the end of the chapter (Table 3.2, Table 3.3, Table 3.4, Table 3.5, and Table 3.6). The pottery recovered from Lillian's Site indicates that the majority of the deposition on the site took place during the Pueblo III period. Pueblo III decorated white ware types include Mesa Verde Black-on-white (N = 42), McElmo Black-on-white (N = 12), Pueblo III White Painted (N = 154), and probably the Late White Painted sherds that are decorated with carbon paint (N = 562); together, the Pueblo III pottery constitutes 84.1 percent of the decorated white ware assemblage. Pueblo II pottery, accounting for 13.2 percent of the decorated white ware assemblage, includes Pueblo II White Painted (N = 11), Mancos Black-on-white (N = 18), and the Late White Painted sherds with mineral paint (N = 92). Pueblo I and Basketmaker III decorated white ware pottery, a total of 2.7 percent of the decorated white ware assemblage, includes Early White Painted (N = 19), Piedra Black-on-white (N = 3), and Chapin Black-on-white (N = 3). Further work is needed to "unmix" the assemblages at Lillian's Site (cf. Kohler and Blinman 1987), but these percentages indicate that the Pueblo III occupation produced the majority of the refuse accumulation at the site.

The Pueblo II occupation appears to have been the next-most-intensive occupation of the site. Fewer decorated white ware sherds are associated with the Pueblo II occupation of Lillian's compared with the Pueblo II occupations at G and G Hamlet and Kenzie Dawn Hamlet. The relative proportion of Pueblo II vs. Pueblo III pottery at Lillian's Site is discussed in Chapter 15.

Although the percentage of decorated white ware sherds dating to the Basketmaker III and Pueblo I periods is small, 223 plain gray, one Chapin Gray, three Moccasin Gray, two Mancos Gray, and four Indeterminate Neckbanded sherds were recovered. The composition of the gray ware assemblage indicates that the Pueblo I occupation was probably more significant than the Basketmaker III occupation. Each of these occupations--Basketmaker III, Pueblo I, and Pueblo II--probably contributed to some of the refuse accumulation at the site.

Complete and fragmentary tools recovered from Lillian's Site include two metates, four manos, eight abraders, two hammerstones, six peckingstones, 30 cores, five projectile points, six bifaces, three drills, two polishing stones, one polishing/hammerstone, 38 modified flakes, 69 modified sherds, one shaped sherd, four bone awls, and 16 pieces of modified bone. A stone pendant fragment was also found. There were 1,983 pieces of chipped-stone debris recovered from the site. Finally, eggshell, mineral samples, and numerous gizzard stones were found.

There is little evidence that the inhabitants of Lillian's Site participated in long-distance exchange networks. San Juan red ware types in the assemblage include Abajo Red-on-orange (N = 2), Bluff Black-on-red (N = 1), and Deadmans Black-on-red (N = 2); these date to the Pueblo I and II periods. Nonlocal red ware also includes five sherds that were so small they could not be assigned to more specific types. Only three stone tools were made from nonlocal material (one piece of obsidian, and two pieces of nonlocal chert/siltstone), and seven of the 1,983 pieces of chipped-stone debris were of nonlocal origin.

Lillian's Site Summary and Interpretations

Lillian's Site is a multiple-component site. Although it was hoped that testing would allow us to identify each component, excavation focused on the Pueblo III occupation. Thirty-five 1-×-1-m sampling units were excavated as a part of the stratified random sample; five judgmentally located 1-×-1-m excavation units were excavated to expand on the randomly located sampling units. The tops of walls were swept and troweled to define the layout of the masonry roomblocks. The presence of masonry roomblocks, a masonry-lined pit structure (kiva), a masonry tower, and a midden dating to the Pueblo III period is the basis for the interpretation that Lillian's Site functioned as a habitation during this time.

Two masonry roomblock units, connected by a curved wall that abuts both, were tested. The roomblocks date to the Pueblo III period. The presence of two roomblock units may indicate that two households resided at Lillian's Site during the Pueblo III occupation. A different, but equally plausible, interpretation is that Lillian's Site was occupied by one large, multigenerational household. The roomblocks may have been partitioned to provide space for these infrahousehold groups (Lightfoot 1994).

A single masonry-lined pit structure, Structure 1, lies in front of the two roomblock units. This structure was probably shared by the occupants of Lillian's Site.

A masonry tower, Structure 4, probably stood at least two stories tall. Structure 4 is located south of Structure 1, so that the southern recess of Structure 1 points at the tower. Structure 4 is also interpreted as a part of the Pueblo III component. This interpretation is based on the masonry construction style and the alignment of Structure 4 relative to the other masonry architectural features.

A low, circular, masonry wall abuts the east side of Structure 4. This construction was built after Structure 4, and therefore also dates to the Pueblo III occupation. This low wall does not appear to enclose a structure and may therefore enclose a plaza.

South of the tower lies the trash mound. This trash mound is aligned with the east roomblock unit, Structure 1, and Structure 4, which suggests that the trash accumulated primarily during the Pueblo III occupation of Lillian's Site. However, pottery from earlier occupations is also present in the midden, indicating that some of the refuse accumulated during these earlier occupations.

Dating the Pueblo III component requires drawing on several lines of evidence. Tree-ring dates for Structure 1 suggest an early A.D. 1200s date, with occupation probably postdating the latest tree-ring date, A.D. 1214 vv. An archaeomagnetic sample from the hearth of Structure 1 yielded a date range of A.D. 1175-1250. Painted white ware pottery associated with the Pueblo III occupation includes both McElmo and Mesa Verde black-on-white types. Wilson (1991), working with pottery from a well-dated site 30 km north of Lillian's Site, dates the beginning of the manufacture of Mesa Verde Black-on-white to after A.D. 1180, and possibly as late as A.D. 1200. Together, the evidence suggests that the Pueblo III occupation of Lillian's Site dates between A.D. 1200 and 1250, and possibly between A.D. 1220 and 1250.

At least one earlier occupation is also represented at Lillian's Site. Collapsed post-and-adobe rooms lie beneath the masonry roomblock. Tree-ring dates for this post-and-adobe roomblock include a date of A.D. 1075 vv. Pottery from these earlier deposits include Mancos Black-on-white, and this supports the tree-ring dates. Basketmaker III and Pueblo I sherds are also present in the deposits below the masonry roomblock; however, the bulk of the evidence suggests that the post-and-adobe roomblock beneath the masonry roomblock dates to late Pueblo II period. Recently excavated sites with post-and-adobe roomblocks located 20 km north of Lillian's Site date between A.D. 1000 and 1100 (Kuckelman and Morris 1988).

We found no evidence of a pit structure associated with the Pueblo II occupation of the site. It may be that our limited testing did not locate this structure, or that Structure 1, the Pueblo III masonry-lined pit structure, was built inside an earlier pit structure, removing all traces of that structure. It is also possible that an earlier pit structure simply is not present. The amount of Pueblo II pottery at Lillian's Site can be compared with the amounts present at G and G Hamlet and Kenzie Dawn Hamlet. At G and G and Kenzie Dawn, both pit structures and surface rooms were associated with the Pueblo II occupation, and these occupations were therefore interpreted as year-round habitations. The fact that Lillian's Site has less Pueblo II pottery than either G and G Hamlet or Kenzie Dawn Hamlet, and the fact that there may not be a Pueblo II pit structure at Lillian's Site, suggests that the Pueblo II occupation at Lillian's did not involve year-round habitation but, rather, a more-limited use of the site, perhaps as a field house.

Pottery made during both the Pueblo I and Basketmaker III times was found at Lillian's Site, with Pueblo I pottery being the more common. It is possible that the site was occupied during these periods, but Basketmaker III and Pueblo I architectural features were not identified. Basketmaker III and Pueblo I sites are present within a .5-km radius of Lillian's Site (Van West et al. 1987; Adler 1988), and the proximity of these sites may have resulted in the mixing of artifacts between them. However, the number of earlier sherds at Lillian's Site favors the interpretation that there was use of Lillian's Site during these early periods. This use probably was limited to certain activities, and did not include use as a habitation during either the Basketmaker III or Pueblo I periods.