The Shields Pueblo Research Project

Annual Report, 2000 Fieldwork

Andrew I. Duff and Susan C. Ryan

Introduction

Shields Pueblo (5MT3807) is a large, Puebloan site with a primary occupation dating between A.D. 1050 and 1300 and an extended occupation that began by at least A.D. 775. Shields is located on property owned by Colorado Mountain College (CMC) and James and Veda Wilson. It is located immediately north of Goodman Point Monument, a unit of Hovenweep National Monument, in Montezuma County, Colorado. This report summarizes the Crow Canyon Archaeological Center's research goals and describes archaeological fieldwork conducted at Shields Pueblo in 2000—our fourth and final field season. Earlier reports describe the archaeological fieldwork conducted in 1997 (Ward 1997), 1998 (Duff and Ryan 1998), and 1999 (Duff and Ryan 2000b). Initial interpretations concerning the occupation and abandonment of Shields Pueblo have been made on the basis of tree-ring dates and the results of preliminary artifact analyses; these interpretations are presented in Duff and Ryan (2000b).

Fieldwork in 2000 was conducted from April through November, under State of Colorado Archaeological Permit number 2000-29. Upon completion of all analyses, an electronic site report on the work conducted at Shields Pueblo will be available on Crow Canyon's Web site.

Research Orientation

Communities through Time: Migration, Cooperation, and Conflict is a regional research design that examines the development and abandonment of ancient Puebloan communities in the Mesa Verde region from A.D. 900 to 1300 and incorporates data gathered at the residential, community, and regional levels. The research conducted at Shields Pueblo was designed primarily to collect artifact and ecofact assemblages from residential structures occupied and abandoned between A.D. 1050 and 1225, the interval during which Shields Pueblo is thought to have been a focal location within the larger Goodman Point community (Adler 1990:260, 1994; Adler and Varien 1994; Varien 1999). The focus of community-level research will involve integrating information from Shields Pueblo into the larger framework of the surrounding natural environment and cultural landscape, defined by the numerous prehistoric settlements surrounding both Goodman Point Pueblo and Shields Pueblo. Regional-level research will compare the Goodman Point community with 26 other, similarly long-lasting, communities in the Mesa Verde region (Varien 1999; Varien et al. 1996), including community centers already investigated by Crow Canyon (Ortman et al. 2000).

Several research questions are oriented toward reconstructing the history of occupation at Shields Pueblo and documenting the changing population levels at Shields and in the immediately surrounding area. These data will allow us to assess the nature and timing of population aggregation into community centers and to evaluate the impact these populations had on their surrounding natural environment. Intensive sampling of wood and charcoal for tree-ring analysis will be used to determine if Shields Pueblo was occupied continuously or if there were intervals during which it was abandoned. These data will also be used to assess the intensity of contemporaneous occupation at Shields during different periods. Tree-ring species information and the results of macrobotanical, palynological, and faunal analyses will be used to reconstruct local environmental conditions and available resources and to evaluate changes in the use or availability of these resources through time.

Within the history of occupation, we are particularly interested in the A.D. 1050–1225 period, the time during which Shields Pueblo appears to have served as a community center for the Goodman Point community. A community center is defined by an area of dense residential and public architecture that was central to a cluster of individual settlements. This span includes the period of Chacoan influence, A.D. 1050–1150, and the post-Chaco period, A.D. 1150–1225. We hope to determine if site organization at Shields and the organization of the Goodman Point community changed from the Chaco to post-Chaco periods and if occupation at Shields was continuous across this transition. Currently, tree-ring dates from Shields indicate use of the location in the A.D. 770s, with a larger and more persistent occupation evidenced by several tree-ring dates from structures spanning the period between A.D. 1100 and 1258 (Table 1). However, there is a gap in the Shields tree-ring record between A.D. 1131 and 1212, which suggests that the site may have been depopulated during this interval (Duff and Ryan 2000a). This gap corresponds to a period of marked environmental deterioration from A.D. 1130 to 1180 (Van West and Dean 2000), and Varien (1999) has suggested that some Mesa Verde-region communities may have been temporarily abandoned during this time.

At about A.D. 1225, the center of the community is thought to have shifted from Shields to Goodman Point Pueblo, a large, aggregated village located just southeast of Shields (Adler and Varien 1994). We initially expected to find that Shields Pueblo had been abandoned at this time. Contrary to our expectations, however, we discovered an estimated 20 residential structures—masonry roomblock-and- kiva complexes—that were constructed and occupied in the mid-A.D. 1200s (Duff and Ryan 2000a, 2000b). What remains to be determined is whether the A.D. 1200s occupation represents continued occupation of Shields or resettlement after a period of abandonment.

The unexpected temporal depth represented at Shields has provided a unique opportunity to evaluate changes to the environment surrounding Shields—both those resulting from regional trends in rainfall and those resulting from human use and alteration of the landscape. The use of various animal species, the deposition of pollen, and patterns in the procurement of construction wood, fuelwood, and gathered resources will be used to differentiate human-induced changes from those resulting from larger-scale environmental processes. Additionally, discrete, well-dated artifact assemblages will permit the evaluation of changes in local exchange, regional exchange, and material procurement throughout the occupation of Shields.

The comparative framework into which data from Shields will be integrated derives from both survey and excavation data. The area surrounding Shields has been surveyed (Adler 1990), with evaluation of surface remains used to estimate the period of occupation and size of recorded sites. Excavation data are available from several sites within the Sand Canyon locality. The Mustoe site, a multicomponent unit pueblo located approximately 1 km southwest of Shields, has been excavated and reported (Gould 1982). Occupation during the Pueblo II and early Pueblo III periods at the Mustoe site provides a local comparative data set for material recovered from Shields. Data from Sand Canyon Pueblo (Bradley 1992, 1993) and from several dispersed settlements surrounding this community center (Varien, ed. 1999) provide information for both direct comparison and an evaluation of differences between these neighboring communities. The late Pueblo III settlement of Castle Rock Pueblo (Kuckelman 2000), also located within the Sand Canyon locality, provides another comparative resource for the late A.D. 1200s occupation at Shields.

As analyses of the artifact and ecofact assemblages collected during the four years of excavation at Shields Pueblo proceed, we will seek to maximize the framework for comparison. We will rely on previous Crow Canyon excavations for which information was collected using the same data recovery and recording protocols as those used at Shields, but we also plan to use data from contemporaneous sites located throughout the Mesa Verde region to place the Shields Pueblo excavations into a regional context.

A Synopsis of Previous Research at Shields Pueblo

Previous work at Shields Pueblo includes unsystematic excavations conducted throughout the mid-1900s. Several vessels recovered from Shields are part of the Chappell collection curated at the Anasazi Heritage Center. A burial uncovered at Shields contained a copper bell, and a short report was published documenting this find (Hayes and Chappell 1962). Colorado Mountain College (CMC) conducted fieldwork at Shields Pueblo as a summer class in archaeology in 1975, 1976, and 1977. CMC excavated portions of five kivas, a roomblock, and other cultural deposits (Bagwell 1975, 1976, 1977). An archaeological survey of the area surrounding Goodman Point and Sand Canyon pueblos was conducted by Crow Canyon researchers in the 1980s, and portions of Shields Pueblo were recorded at that time (Adler 1990, 1992).

Eighteen high-density concentrations of artifacts and sandstone rubble were identified during survey and mapping in 1996 (Figure 1). Crow Canyon excavations at Shields Pueblo began in 1997. A staged research approach for the investigation of Shields was developed. Research in each high-density area began with systematic surface collection and the excavation of a randomly selected sample of subsurface deposits using 1-x-1-m units. These excavations were followed by the judgmental sampling of cultural deposits and structures. The 1997 season was devoted to obtaining a systematic sample from, and assessing the integrity of, deposits in the southern half of the site (Ward 1997). A remote-sensing survey conducted at the end of the 1997 field season has been used to identify the probable locations of buried architecture (Varien 1997). The 1998 field season was devoted to furthering our systematic sample and to the archaeological testing of possible subterranean structures identified by remote sensing, almost all of which proved to be buried prehistoric structures (Duff and Ryan 1998; Duff et al. 1999). During the 1999 field season, we expanded our sample of structures through extensive testing of remote-sensing anomalies, and we retrieved artifact and ecofact assemblages from structures and deposits representing the full range of occupation at Shields (Duff and Ryan 2000b). Approximately 35 structures had been tested at Shields by the end of the 1999 field season. Excavation had documented a Pueblo I- period (A.D. 700–900) presence at the site, several late Pueblo II-period (ca. A.D. 1050–1150) structures and deposits, and an initially unexpected, but extensive, late Pueblo III-period (A.D. 1225–1300) occupation. The extent of occupation during the early Pueblo III period (A.D. 1150–1225) remains uncertain. A detailed summary of the work conducted during each of the first three field seasons and preliminary interpretations of the available data are presented in Duff and Ryan (2000b).

Research Goals, 2000 Season

Crow Canyon's goals for the 2000 field season were as follows: (1) to complete the collection of surface artifacts from a 3-m-diameter portion of each 20-x-20-m block within the site; (2) to complete the systematic sample of deposits at Shields by excavating randomly selected 1-x-1-m test units in Blocks 1600, 1700, and 1800; (3) to complete excavation and documentation of all unfinished units; and (4) to ensure the recovery of assemblages and samples from deposits and structures dating to each period of occupation at Shields. The last goal was accomplished through further excavation of already exposed structures and deposits and the testing of additional remote-sensing anomalies. The remainder of this report documents the work conducted to achieve these goals.

Field Methods, 2000 Season

Collection of surface artifacts and the excavation of randomly selected 1-x-1-m units was conducted in areas of the site that had not been previously investigated. We also conducted surface collections in areas of the site surrounding each of the 18 high-density blocks. Table 2 lists the center point of each surface collection completed this year. Our primary excavation strategy involved the judgmental testing of remote-sensing anomalies and the excavation of portions of structures exposed during this process. Remote-sensing anomalies were initially tested by means of an east-west trench across the anomaly excavated through plow-zone-disturbed and natural deposits (see Duff and Ryan 1998; Duff et al. 1999). Trenching was used to define structure boundaries and to reveal the sequence of deposition in structures. Hand excavation was then used to recover artifact assemblages from structure fill, floor, and feature contexts. Structure testing was concentrated in areas of the site from which we had not yet investigated remote-sensing anomalies. However, we also tested remote-sensing anomalies in areas of the site with already-exposed structures dating from the Pueblo II period (ca. A.D. 900–1150) in an effort to expand the sample of structures from this period. In addition, excavations in several structures and deposits identified during previous seasons were expanded to secure large and systematic samples. The excavation of 2-x-2-m units was conducted in areas of high artifact density associated with tested structures. These judgmental units were generally placed southeast of known structures where midden deposits typically occur (Lipe 1989) and were excavated entirely by hand. The methods employed in 2000 represent a continuation of those applied in previous seasons, and these are described in greater detail in Duff and Ryan (1998, 2000b) and Duff et al. (1999).

Field Personnel, 2000 Season

The following people contributed to Shields Pueblo fieldwork in 2000: Andrew Duff, project director; Susan Ryan, assistant project director; Melissa Churchill, Drew Kenworthy, and Hugh Robinson, research archaeologists; Mark Varien, director of research; Melinda Bernardo, Diane Curewitz, Devin White, and Darcy Wiewall, research interns; Jane Dillard, research volunteer; the Crow Canyon education staff; and the Crow Canyon program participants who made the work accomplished this year possible.

Excavations, 2000 Season

Excavation during the 2000 season was concentrated within 10 of the 18 high-density block areas; portions of Blocks 100, 200, 400, 1100, 1200, 1300, 1400, 1600, 1700, and 1800 were examined this season. Randomly selected 1-x-1-m test units were excavated in Blocks 1600, 1700, and 1800. The excavation of structures and midden deposits was concentrated in the Block 100, 200, 400, 1100, 1200, 1300, and 1400 areas. Table 3 lists each of the excavation units by block and presents information about the size and average depth of each unit, whether features or structures were found, and whether the unit was completed at the end of the field season. All ground-disturbing activities at Shields Pueblo have been completed, and no additional excavations are planned. As can be seen in Table 3, excavation in several units was terminated before sterile deposits were reached. In most cases, excavation was halted because we had obtained sufficient material to satisfy the requirements of the research design; in a few cases, time constraints precluded complete excavation. Such cases are differentiated in the following descriptions.

Figure 1 illustrates the locations of all excavations conducted at Shields. Excavations, grouped by block, are discussed in greater detail below. Dating and research summaries for the Block 300, 500, 600, 700, 800, 900, and 1000 areas were detailed in Duff and Ryan (2000b), and no additional information for these areas was obtained this season. The current report documents areas investigated this season and previously reported structures for which new data—primarily, tree-ring dates—have since been received. The time during which a structure or area was used was inferred on the basis of decorated pottery types and/or tree-ring dates. Dating of the pottery assemblages follows Wilson and Blinman (1991).

Block 100

Work in Block 100 was designed to complete or expand the excavation of structures exposed in previous years. Figure 2 plots the locations of units and structures examined.

Nonstructure 101 and Nonstructure 148

A possible subterranean room in the northern portion of the Structure 110 area was investigated with a single 1-x-2-m unit (478N 488E). A small portion of a subterranean masonry wall suggested the presence of an underground room, but this unit was not excavated to sufficient depth to definitively determine the extent and function of the wall. Instead, cultural fill (Nonstructure 101) above the wall was excavated, which will help date its construction. The wall was recorded as a feature (Nonstructure 148, Feature 1), though it is probably associated with an underground room.

Structure 122

Structure 122 is a Pueblo II kiva that was burned at abandonment, filled with trash, and then had a smaller kiva (Structure 123) constructed within it (Duff and Ryan 2000b). Portions of Structure 122 and almost all of Structure 123 were excavated by CMC in 1977 (Bagwell 1977). An excavation unit in intact deposits located between the two structures was completed by Crow Canyon in 2000. Burned primary beams from the collapsed roof of the larger kiva, as well as a small portion of the kiva floor, were found in this test unit. Tree-ring samples from burned, upper roof fall include a cutting date of A.D. 1114 and a number of noncutting dates that extend to A.D. 1124 (Table 1). Additional tree-ring dates are likely to derive from samples collected in 2000.

Structure 124

Structure 124 is an earthen-walled, subterranean room that was used for corn grinding. Two mealing bins, the tops of which were exposed in 1999, and an associated pit were excavated during the final field season. The upright sandstone slabs defining the grinding bins remained, but the mealing stones (metates) themselves had been removed before the structure was abandoned. An adjacent pit feature, the function of which could not be determined, was also excavated. This pit feature may be associated with the floor of Structure 124, and it may have been used for storage.

Structure 138

Structure 138 is a pit structure uncovered during the excavation of Backhoe Trench 128. The southern recess of this structure was defined in the bottom and profile of the trench. A 1-x-2-m unit excavated within this structure in 1999 (Segment 8) defined its southern edge and ventilator tunnel but did not reach the floor. A 2-x-2-m unit (Segment 5) was placed immediately north of the original unit this season, and both segments were excavated to the structure floor. Structure 138 proved to be a relatively small, deep, earthen-edged pit structure with four pilasters and no bench. The southern recess pilasters were earthen, whereas the northern pilasters—only one of which was preserved—were masonry. A hearth and an ashpit were excavated, and ash samples were collected from both for a variety of analyses. It appears that the roof timbers were removed, and the roof collapsed. The structure was later filled with midden. The types of pottery recovered from the trash fill suggest that the structure probably predates A.D. 1140.

Structure 139

The southern recess and portions of the vent shaft of Structure 139 were excavated in 1999; in 2000, Backhoe Trench 147 was excavated to expose the main chamber. A rectangular unit (Segment 1) was excavated to the floor, exposing the hearth and an ashpit. Ash from each feature was sampled for analysis. Structure 139 is similar in construction to Structure 138, having earthen edges and earthen southern recess pilasters. One masonry pilaster was exposed, and it is unclear if Structure 139 had four or six roof-support pilasters. The roof of Structure 139 appears to have been salvaged at abandonment, and the depression then filled in naturally. The pottery assemblage and the architectural characteristics of this structure suggest that it was constructed sometime before A.D. 1150.

Structure 140

Structure 140 is a wattle-and-daub surface room that was uncovered in the western end of Backhoe Trench 128. A tree-ring sample from this structure produced a cutting date of A.D. 1103, the earliest Pueblo II cutting date recovered at Shields so far. Portions of the hearth and structure floor were excavated in 1999. This season, a 2-x-1-m unit (Segment 10) exposed more of the wattle-and-daub wall north of the trench. This segment exposed six additional burned posts, each of which was encased in a burned adobe footing. Each post was collected for tree-ring dating.

Structure 149

Structure 149 was identified in the profile of Segment 7, a 2-x-2-m unit completed this season. Structure 149 encompasses most of Segment 7, and the bulk of the deposits excavated from this context appear to be fill within the structure. The hearth was exposed in profile, and ash was sampled for analysis. After abandonment, the room was filled with trash. Two pit features were excavated through this fill, extending below Structure 149. A wood sample from this trash deposit produced a near-cutting tree-ring date of A.D. 1125 (Table 1), previously listed as a date associated with Arbitrary Unit 105 (Duff and Ryan 2000b:Table 2). The structure may have been related to the adjacent Block 100 roomblock (Structures 102, 103, 104, and 121), as it appears to have been constructed on approximately the same surface and the tree-ring date suggests contemporaneity.

Structure 150

Structure 150 was exposed in Backhoe Trench 128 and was recorded in profile only. The trench appears to have exposed the southern recess of Structure 150, and the structure appears to have been filled with trash after abandonment. The structure is located immediately south of the Block 100 roomblock, suggesting contemporaneity, which would place use of this structure in the early A.D. 1100s.

Block 200

Work in the Block 200 area this season included the further exploration of already-uncovered features and structures, the testing of one additional remote-sensing anomaly, and the excavation of four 2-x-2-m judgmental units. Figure 3 shows the locations of all of the Block 200 excavation units.

Nonstructure 210

Nonstructure 210 is a midden deposit that was initially exposed in a randomly selected 1-x-1-m unit (517N 523E). A second 1-x-1-m unit (516N 523E) adjacent to the first was completed this season. A small section of wall uncovered in the southwest corner of the unit was recorded as a feature (Nonstructure 210, Feature 1). Midden deposits continued in the portion of the unit bounded by the wall, with sterile soil exposed in the remainder of the unit. This feature appears to have been part of the upper lining wall of a kiva. The results of pottery analysis suggest that the midden was deposited sometime prior to A.D. 1200.

Structure 221

Excavation of Structure 221, a burned masonry kiva, was completed last season. Additional tree-ring dates were received for this structure, and these include a cutting date of A.D. 1248—a few years later than the previous latest date from this structure (Table 1). Structure 221 appears to have been constructed in the A.D. 1240s, and the fact that its roof beams were not salvaged suggests that it was one of the last structures to be abandoned at Shields.

Structure 223

Structure 223 is a masonry-lined kiva with burned roof-fall deposits, although primary roofing elements appear to have been removed prior to abandonment. This season, a small segment was excavated to reveal a portion of the hearth. The northern end of the hearth was excavated, and ash was collected for analysis. Several tree-ring dates from Structure 223 indicate that it was used into the A.D. 1250s (Table 1).

Structure 224

The excavation of Structure 224, an intensively burned masonry kiva, was completed this season. An east-west cross section of this kiva was excavated in two segments, exposing the floor and a hearth. Portions of at least two complete vessels were recovered from the hearth area, and ash from the hearth was collected for analysis. Several hundred pieces of burned wood were recovered, and additional tree-ring dates were obtained from analysis of samples collected last season (Table 1). These indicate that Structure 224 was used into the A.D. 1250s, and the presence of primary beams suggests that it was among the last structures to be abandoned at Shields.

Nonstructure 233

Nonstructure 233 is a midden that was deposited in Structure 234 after that structure had fallen out of use, the roof had collapsed, and it had almost completely filled with natural deposits. Nonstructure 233 was investigated by means of a single 2-x-2-m unit (504N 546E) that began at modern ground surface, continued below the plow zone, and was terminated when the natural fill within Structure 234 was exposed. Nonstructure 233 appears to have been deposited in the A.D. 1200s and may be an extension of the deposit noted in several 2-x-2-m units north of this area (Nonstructure 238, see below).

Structure 237

An east-west segment in Structure 237, an earthen-edged pit structure without a bench, was completed in 1999. This unit was excavated through roof collapse, and it exposed the structure floor and a hearth. The upper fill consisted of midden deposited into the structure after it had collapsed. In 2000, a 2-x-2-m unit abutting the original trench on the south was excavated beginning at modern ground surface, providing a sample of the midden deposited into Structure 237. This unit was terminated when roof-fall deposits were uncovered. The results of pottery analysis suggest that the midden was deposited during the late Pueblo II period.

Nonstructure 238

Nonstructure 238 is an extensive midden deposit preserved below the plow zone, south of Structures 221, 222, and 223. This deposit was identified in three different 2-x-2-m units in 1999, and each was completed this season. The boundary of this deposit was expanded to include a unit (508N 550E) south of Structure 224. One of the original units (510N 540E) was expanded using another 2-x-2-m unit (509N 541E) to further examine a possible feature. Portions of two coursed-masonry walls and several upright slabs were uncovered in these two units, and each was recorded as a feature. Additional upright slabs in the adjacent unit (511N 544E) were recorded as a feature, as was a segment of a masonry wall identified in 508N 550E. The orientations of these features did not suggest an identifiable structure, but it is likely these features represent basal courses associated with the masonry walls of one or more structures. The upper courses of these walls were dismantled, and then midden was deposited across this area. Several of these features had been constructed on top of, or had been excavated into, sterile sediment, but one wall segment (508N 550E) was constructed atop cultural fill. The location of these features suggests that they may have been remnants of a surface structure associated with Structure 234, a deep pit structure located south of these grid units. Additionally, midden constituting the uppermost fill in Structure 234 (Nonstructure 233) may represent a continuation of the Nonstructure 238 deposit.

Nonstructure 239

Nonstructure 239 is a midden deposit preserved below the plow zone, south of Structure 225. It was exposed in a 2-x-2-m unit (512N E565E). Only a few centimeters of this deposit were preserved, but a pit feature was identified and excavated. The function of the pit is uncertain, but it may have been used for extramural storage.

Structure 241 and Nonstructure 242

Structure 241 and Nonstructure 242 were exposed in Backhoe Trench 240, which was excavated to test a remote-sensing anomaly. Structure 241 is a burned, masonry-lined kiva. An L-shaped unit (Segment 1) was excavated to the floor, exposing a wall niche, a pilaster, a sipapu, and the hearth. Numerous primary roofing elements were recovered, and ash from the hearth was sampled for analysis. The artifacts recovered from the floor and the presence of primary roofing elements suggest that Structure 241 was used into the A.D. 1250s.

Nonstructure 242 consists of two separate masonry wall segments (Features 1 and 2) exposed during the excavation of Backhoe Trench 240, and these were recorded in profile only. These wall segments occur above the roof fall of Structure 241 but are within the pit excavated for Structure 241. Feature 1 is a wall segment with several courses of masonry that may have been constructed across the southern recess of Structure 241. Feature 2 has a maximum of three courses of masonry and is located immediately atop roof fall above a masonry pilaster of Structure 241. These two features resemble post-abandonment structures noted elsewhere at Shields in that they occur above burned, late-abandoned, masonry kivas (see Duff and Ryan [2000b], Ryan [2000], and the Block 400 description below). However, the exposed portions of Nonstructure 242 do not appear to have been part of a circular construction similar to others noted at the site. The function of Nonstructure 242 is unknown, although these walls clearly postdate the collapse of Structure 241.

Structure 243

A kiva completely excavated by CMC in 1975 was designated Structure 243, although no excavation was conducted in this structure by the Crow Canyon Archaeological Center. The kiva is located in Block 200 and was relocated last season using CMC excavation documents (Bagwell 1975). Wood from this structure produced cutting dates of A.D. 1129 and 1131 (Table 1).

Block 200 2-x-2-m Units

Four 2-x-2-m units were excavated to collect assemblages associated with Block 200 structures; two were located southeast of Structure 237, and two were located southeast of Structure 241. Each exposed sterile sediment immediately below the plow zone, with no features or intact deposits uncovered.

Block 400

Portions of the Block 400 area were excavated by CMC in 1976 and 1977. Using CMC excavation documents (Bagwell 1976, 1977), we were able to relocate the areas CMC investigated, and information about this work is reported below. We have assigned Crow Canyon study unit numbers to many of the CMC excavations, even though Crow Canyon did not excavate in any of these structures. Four randomly selected 1-x-1-m units were excavated in the Block 400 area in 1997 (Ward 1997). This season, portions of two pit structures and four midden units were excavated. The locations of all Block 400 area excavations are plotted on Figure 4.

CMC Excavations in the Block 400 Area (Structures 408, 409, 410, and 411)

During the 1976 field season, a backhoe trench, a small masonry room, a kiva, and twelve 10-x-10-ft squares were excavated during a CMC field course (Bagwell 1976). The structures excavated by CMC are plotted on Figure 4, but their locations should be viewed as approximate. Excavation began with a 30 yard east-west backhoe trench (Backhoe Trench 412) designed to locate subterranean structures. This trench uncovered a small masonry room (Structure 411) that CMC then partially excavated. Directly north of this room, and connected to it by an unlined tunnel, was a masonry kiva (Structure 408; CMC Feature A, 1976). Structure 408 was divided into quadrants and excavated in three levels: upper fill, consisting of fill above the pilasters; lower fill, consisting of fill from the top of the pilasters to four inches above the floor; and floor fill, consisting of the last four inches of fill above floor and floor artifacts (Bagwell 1976). Shortly after excavation of the upper fill in Structure 408 began, CMC exposed a circular, coursed-masonry wall. This was designated Feature B (Structure 409) and interpreted as a "tower." Structure 409 was photographed and then removed by CMC to facilitate the excavation of the kiva (Structure 408). CMC excavation notes make reference to burned roof fall lying directly on the structure floor (Bagwell 1976), but no tree-ring dates were obtained. Floor features included a masonry deflector with a niche, a circular hearth, and a sipapu. Based on structures with similar abandonment and fill sequences, Structure 408 is likely to have been used into the mid-A.D. 1200s. Structure 409 clearly postdates the abandonment of Structure 408, and appears to be a shrine-like construction built as Shields was abandoned (Duff and Ryan 2000a, 2000b; Ryan 2000). CMC also excavated 12 10-x-10-ft units in the area south of Structure 408 in an attempt to locate midden deposits.

During the 1977 field season, another kiva (Structure 410, CMC Feature C 1977) was excavated in the Block 400 area (Bagwell 1977). Structure 410 was located approximately 20 ft west of Structure 408. Records do not indicate how the structure was excavated, but it is assumed that the methods described above were employed. The northeast quadrant was left unexcavated, providing a stratigraphic balk. The upper lining wall of Structure 410 was earthen except for the south wall of the southern recess, which was masonry-lined. Floor features included a circular hearth, a rectangular ashpit, a coursed- masonry deflector, a bin of upright slabs located under the eastern pilaster, and two pairs of shallow circular holes set into the floor (one pair in the northwest quadrant, the other in the southwest quadrant). Structure 410 also contained an unlined, bench-top tunnel located in the northwest quadrant between two masonry pilasters. The tunnel appeared to veer to the northwest, possibly connecting the kiva to associated surface rooms. Excavation notes make reference to burned roof fall lying directly on the floor. Wood from this structure produced several tree-ring dates (Table 1), the latest of which is a cutting date of A.D. 1212.

Structures 405, 407, and 401

Backhoe Trench 404 was excavated in 2000 to test a remote-sensing anomaly. This trench exposed three different structures: Structure 405, a burned, masonry-lined kiva; Structure 401, a circular masonry structure located on fill above the upper lining wall of Structure 405; and Structure 407, a circular masonry structure constructed atop the natural, postabandonment fill within the depression created by the collapse of Structure 405.

Two segments were used to examine Structure 405, the earliest, stratigraphically, of the three structures. Segment 1 was excavated in the north-central part of the structure. Excavation proceeded through a natural strata with abundant sandstone rubble and burned roof fall until the floor was exposed. A sipapu was exposed and excavated, and the northern portion of the hearth was uncovered. Excavation of the hearth revealed a clean silty fill, a sample of which was collected for analysis. It appears that the hearth had been cleaned of ash prior to the burning and collapse of Structure 405. Segment 2 exposed a portion of the upper lining wall and bench surface along the east edge of Structure 205.

Structure 407 is a circular stone structure constructed atop natural deposits that accumulated after the collapse of underlying kiva, 405. was dry-laid, blocky-sandstone masonry, and it ranged from three to seven courses in height. interior face this construction exposed Segment 1, exterior 2. No features associated with were detected, not on formal surface.

Structure 401 is another circular stone structure constructed on deposits that had accumulated naturally above the collapsed roof of Structure 405. In this case, however, the wall of the later structure was built on top and outside of the upper lining walls of Structure 405. Structure 401 consists of three to four courses of shaped, tabular and blocky sandstone set in mortar. Structure 401 is the only recorded circular masonry structure found atop a burned kiva at Shields that has a diameter exceeding that of the underlying structure. It is otherwise similar to many other circular constructions, e.g., Structure 407, that were built atop burned, late-abandoned kivas at Shields (see Duff and Ryan 2000b).

The function of Structures 401 and 407 is unknown, but they may have been shrine-like features constructed as people left the site and the region (Duff and Ryan 2000a; Ryan 2000). Kivas elsewhere at Shields with circular stone structures constructed atop collapsed kiva roofs postdate A.D. 1245 (see Duff and Ryan 2000b), and it is likely that Structure 405 also postdates A.D. 1245. Structures 401 and 407 postdate the abandonment of Structure 405.

Structure 406

Backhoe Trench 403 was excavated to test a remote-sensing anomaly, and it exposed a burned, masonry-lined kiva (Structure 406). The initial east-west trench was expanded to the south to ensure exposure of the hearth. An L-shaped unit (Segment 1) was excavated to the floor of Structure 406. Fill consisted of plow zone, a dark organic stratum below the plow zone, natural postabandonment fill, and burned roof fall from which numerous wood samples were collected for tree-ring analysis. The primary roofing elements appear to have been removed prior to the burning of Structure 406, and a piece of ground stone was the only artifact left on the portion of the floor exposed. A square, slab-lined hearth was excavated, and ash was collected for analysis. Structure 406 appears to have been used into the late Pueblo III period but appears to have been abandoned while portions of the site continued to be occupied.

Block 400 2-x-2-m units

Four judgmental 2-x-2-m units were excavated in the Block 400 area to recover assemblages associated with Structures 405 and 406. Two units were located southeast of each structure, and all four units exposed sterile deposits immediately below the plow zone.

Block 1100

Twelve 2-x-2-m units, four backhoe trenches, and portions of four structures were excavated in the Block 1100 area this season. Surface collections and the excavation of four 1-x-1-m units and five 2-x-2-m units were conducted in 1998 (Duff and Ryan 1998). The locations of all excavation units are plotted in Figure 5.

Structure 1106

Backhoe Trench 1104 was excavated to test a remote-sensing anomaly, and it exposed Structure 1106, a masonry-lined kiva. A masonry pilaster and upper lining wall were exposed in the western portion of the trench, and a masonry upper lining wall was exposed in the east. Backhoe Trench 1104 was expanded to the north to ensure exposure of the hearth. An L-shaped unit (Segment 1) was excavated until the floor of Structure 1106 was exposed. Fill consisted of disturbed sediments (plow zone), a natural dark organic stratum below the plow zone, natural postabandonment fill, and burned roof fall from which numerous wood samples were collected (including primary beam fragments) for tree-ring analysis. A square hearth was excavated, and ash was collected for analysis. The presence of primary beams and floor artifacts suggests that this structure is likely to have been used into the mid-A.D. 1200s.

Nonstructure 1107 and Structure 1108

Backhoe Trench 1105 was excavated to test a remote-sensing anomaly, and intact midden deposits (Nonstructure 1107) were uncovered directly below the plow zone. Sterile sediments were defined on the eastern and western portion of the trench. Excavation within the trench (Segment 1) continued through the midden deposit, exposing the earthen walls of Structure 1108, a kiva. A 2-x-2-m unit (656.65N 375.5E) was begun at modern ground surface north of the trench to sample midden deposits (Nonstructure 1107) and to uncover the hearth in Structure 1108. Excavation of the 2-x-2-m unit continued until the floor of Structure 1108 was exposed. The north half of a circular hearth was excavated, and ash was collected for analysis. Fill consisted of plow zone, intact midden (Nonstructure 1107), natural fill, and unburned roof fall. Although small pieces of wood were present in the roof fall, it appears that the primary beams were salvaged at the time of abandonment.

In 1998, two midden units were excavated a few meters to the north and east of the Nonstructure 1107/Structure 1108 excavations, uncovering a burned, slab-lined pit feature (Nonstructure 1102). This feature contained wood that produced a noncutting date of A.D. 1199, and it may have been contemporaneous with the deposition of Nonstructure 1107 midden. Structure 1108 appears to have been constructed earlier than other structures investigated in Block 1100, and it was probably used sometime in the early A.D. 1100s.

Nonstructure 1109

A 2-x-2-m unit (628N 362E) was placed southeast of Structure 1106 to collect associated midden deposits. Intact midden deposits (Nonstructure 1109) were uncovered below the plow zone across the entire unit, and excavation continued until sterile sediments were reached. Additional 2-x-2-m units (628N 364E and 630N 362E) were placed adjacent to the original unit on the north and east to further define the boundary of Nonstructure 1109 and to collect associated assemblages. Both units contained intact midden deposits below the plow zone, and excavation continued until sterile sediments were reached. All three units revealed midden deposited atop an irregular surface resulting from prehistoric excavation. Nonstructure 1109 probably served as a borrow pit prior to being filled with midden.

Structure 1113

Backhoe Trench 1110 was excavated to test a remote-sensing anomaly, and it exposed Structure 1113, a masonry kiva. Excavation within the eastern two-thirds of the trench (Segment 1) was conducted until the floor of Structure 1113 was exposed. Fill consisted of disturbed sediments, a dark organic stratum below the plow zone, natural postabandonment fill, and unburned roof fall. The roof timbers appear to have been salvaged at the time the structure was abandoned. A greatly deteriorated ashpit, from which ash had been removed prior to abandonment, was exposed in the southwestern portion of Segment 1. A circular, remodeled hearth was uncovered directly northwest of the ashpit, and ash from the hearth was collected for analysis. Structure 1113 appears to have been constructed and used in the Pueblo III period.

Structure 1114

Backhoe Trench 1111 was excavated to test a remote-sensing anomaly, and it revealed a masonry kiva (Structure 1114). A masonry pilaster and upper lining wall were exposed in the eastern portion of the trench, and a masonry upper lining wall was exposed in the west. Excavation in the eastern two-thirds of the trench (Segment 1) was conducted until the floor of Structure 1113 was exposed. Fill consisted of plow zone, a dark organic stratum, natural postabandonment fill, and burned roof fall from which numerous wood samples were collected (including primary beam fragments) for tree-ring analysis. The northern portion of a circular hearth was excavated, and ash was collected for analysis. Based on the structure's architecture and the floor assemblage, it appears that Structure 1114 was constructed during the late Pueblo III period.

Nonstructure 1116

A 2-x-2-m unit (649.4N 362.6E) was placed southeast of Structure 1114 to sample associated midden deposits. A feature or possible structure (Nonstructure 1116) was uncovered in the western portion of the unit at the base of the plow zone. Sterile sediment was exposed in the eastern portion of the unit. Excavation within Nonstructure 1116 revealed the edge of a circular pit and exposed natural strata with silty laminations atop what appeared to be burned roof fall containing abundant burned corn. Excavation ceased before the nature of the deposits and the feature could be determined, but the burned material suggests the likelihood of a pit structure.

Block 1200

Five 2-x-2-m units, three backhoe trenches, and portions of three structures were investigated in the Block 1200 area this season. The locations of these units, and those of four previously excavated 1-x-1-m units, are plotted on Figure 6.

Nonstructure 1202

A randomly selected 1-x-1-m unit (594N 355E), excavated in 1998, uncovered intact midden deposits (Nonstructure 1202). This season, two 2-x-2-m units (595N 355E and 597N 355E) were excavated to the north of the original unit to further sample and define this deposit. Excavation ceased when an irregular sterile surface was exposed. The irregular surface appears to have been the base of a prehistoric borrow pit into which midden was later deposited. The results of pottery analysis from the midden suggest that it was deposited during the Pueblo III period.

Structure 1205

Backhoe Trench 1203, excavated to test a remote-sensing anomaly, revealed a masonry kiva (Structure 1205). A masonry pilaster was exposed in the western portion of the trench as was a masonry upper lining wall in the east. Excavation within the western two-thirds of the trench (Segment 1) continued until the floor was exposed. Fill consisted of disturbed sediments (plow zone), a dark organic stratum below the plow zone, natural postabandonment fill, and burned roof fall. Numerous wood samples, including primary beam fragments, were collected for tree-ring analysis. The southern portion of a circular hearth was exposed, and ash was collected for analysis. Based on the architecture, the floor assemblage, and the presence of primary beams, Structure 1205 was probably constructed and used in the mid-A.D. 1200s.

Structure 1206

Backhoe Trench 1204 was excavated to test a remote-sensing anomaly and revealed Structure 1206, a masonry kiva. The original east-west trench uncovered portions of the south edge of the structure near the southern recess; therefore, the trench was expanded to the north to ensure access to the hearth. Excavation of a rectangular unit (Segment 1) near the center of the structure exposed the floor and a circular hearth from which ash was collected for analysis. Fill consisted of disturbed sediments, a dark organic deposit below the plow zone, natural postabandonment fill, and unburned roof fall. The roof timbers appear to have been salvaged at abandonment. Structure 1206 was probably constructed and used during the Pueblo III period.

Nonstructure 1207

Two 2-x-2-m units (602N 347E and 604N 347E) were excavated in the area south of Structure 1206 to sample midden deposits associated with this structure. Cultural fill was uncovered below the plow zone in both units (Nonstructure 1207). To identify the boundaries of, and the nature of the deposition in, Nonstructure 1207, Backhoe Trench 1208 was excavated from the base of excavation within the 2-x-2-m units along their east margin. The trench was excavated until sterile deposits defined the edge of a pit. This occurred within the southern 2-x-2-m unit (602N 347E), but the trench extended north beyond the units. The trench revealed a large, irregular pit that may have been intended to become a pit structure, but it had been abandoned prior to completion. The fill sequence consisted of plow zone, a natural dark organic stratum, and what appeared to be a redeposited, mixed fill with a heavy caliche content that extended to the limits of prehistoric excavation.

Structure 1209

A 2-x-2-m unit (603N 363E) placed southeast of Structure 1205 to recover associated midden deposits uncovered a portion of a circular, earthen-walled pit structure (Structure 1209). The walls of Structure 1209 were unlike those of other earthen pit structures investigated at Shields Pueblo in that they sloped out substantially from top to bottom, similar to a bell-shaped pit. Fill in the structure consisted of disturbed sediments, natural postabandonment fill, and unburned roof fall. The portion of the floor exposed suggests that no bench was present. The function of this structure is unknown, but the presence of a roof and a prepared floor indicates a formal subterranean room.

Block 1300

In 2000, work within the Block 1300 area included the expansion of operations in areas initially examined last season, the testing of two remote-sensing anomalies, the further excavation of exposed structures, and the completion of several other grid-based units (Figure 7).

Nonstructure 1303

Nonstructure 1303 is an intact cultural deposit uncovered in a randomly selected 1-x-1-m unit (543N 642E) in 1998. This unit is located northwest of Nonstructure 1309, immediately north of Backhoe Trench 1304. Sterile sediments were exposed in the south half of this unit, and excavation was terminated, although fill continued. Pottery in this deposit includes both Pueblo II and Pueblo III types.

Structure 1307

Structure 1307 was examined in 1999. It consists of a partly subterranean, earthen-edged, circular pit structure that was later remodeled into a smaller structure. Upon abandonment, the structure was filled with midden deposits. Two 2.4-x-2-m units abutting the trench to the north were used to sample this midden deposit, one of which was also excavated to the floor of the structure. No additional floor features were uncovered. The results of pottery analysis indicate that Structure 1307 dates to the Pueblo I period and that the overlying midden was deposited during the Pueblo II period.

Structure 1308

A portion of Structure 1308 had been excavated to floor last season, sampling structural fill and revealing an earthen-walled pit structure with a bench. The fill sequence in Structure 1308 consisted of plow zone, an artifact and ash-rich midden deposit, some intermixed cultural and natural deposition, and unburned roof collapse. A 2.3-x-2-m unit was excavated south of the original trench this season, beginning at modern ground surface and continuing until the floor of Structure 1308 was exposed. This unit provided a thorough sample of the midden deposit. A circular hearth that had been remodeled several times and a probable ashpit were revealed when the floor was exposed. Ash from these features was sampled for analysis. Structure 1308 appears to date to the Pueblo II period, and the overlying

Nonstructure 1309

Nonstructure 1309 was originally thought to be a structure on the basis of what was exposed within Backhoe Trench 1304 last season, but excavation of two units south of the trench in 2000 determined it to be a trash-filled borrow pit. The study unit type has been updated. A 1.9-x-2-m unit (540N 644.5E) and 1-x-2-m unit (539N 644.5E) were excavated until an uneven, sterile surface was exposed. Fill consisted of plow zone and a midden deposit that contained abundant sandstone. The results of pottery analysis suggest that this midden was deposited in the early A.D. 1100s.

Nonstructure 1310

In 1999, the plow zone was mechanically removed from a 10-x-12-m area to expose potential structures suggested by remote-sensing anomalies. A large cultural deposit encompassing most of the west half of this stripped area was thought to be the outline of a pit structure, and this deposit was investigated with several units this season. Four 2-x-2-m units and one 1-x-2-m unit were excavated into this deposit, revealing an ash-rich midden deposited into a borrow pit. An uneven sterile surface was uncovered throughout this area. Nonstructure 1310 appears to have had soil removed for use elsewhere, to have filled in naturally with some limited cultural deposition, and then to have been used for trash deposition. The midden appears to be a Pueblo II period deposit.

Nonstructure 1311

In the eastern portion of the 10-x-12-m mechanically stripped area, a cultural deposit lacking clear boundaries was identified. Two 2-x-2-m units were excavated into this deposit, both of which uncovered sterile sediment within 20 cm. The deposits in Nonstructure 1311 are of uncertain origin, but they resemble the mixed natural and cultural deposit found at the base of the adjacent Nonstructure 1310, suggesting that Nonstructure 1311 may also have been a borrow pit.

Nonstructure 1312

Nonstructure 1312 is a cultural deposit preserved below the plow zone, which was uncovered in two 2-x-2-m units (532N 627E and 532N 629E). The fill of Nonstructure 1312 contains artifacts, but artifact density is lower than in typical midden deposits. The disturbed upper fill (plow zone) in these units was characterized by high artifact density, and Nonstructure 1312 may have been another borrow pit that was later filled with midden.

Structure 1315

Backhoe Trench 1313 was excavated to test a remote-sensing anomaly, and it exposed a burned, masonry-lined kiva (Structure 1315). A single segment (Segment 1) was excavated in the east half of the structure from the base of the trench to the structure floor. The fill in Structure 1315 consisted of plow zone, a dark organic stratum, natural deposits, and burned roof fall. Several primary roof beams were sampled for tree-ring analysis. A circular hearth was exposed and excavated, and ash was sampled for analysis. An intrusive feature, excavated through burned roof fall, rested on the structure floor. This feature was slab-lined, but its function is unknown. Structure 1315 appears to have been constructed and used during the late Pueblo III period.

Structure 1316

Backhoe Trench 1314 was excavated to test a remote-sensing anomaly, and it revealed Structure 1316, a burned, masonry-lined kiva. The initial trench exposed the southern recess and was expanded north to facilitate excavation of the central portion of the structure. Structure 1316 was excavated in two segments. Segment 1 exposed the northern portion of the southern recess. Segment 2, a large rectangular unit, exposed the north edge of the structure, a portion of the bench, a masonry pilaster, a niche in the bench face, a sipapu, and an elliptical hearth. Each feature was excavated, and ash from the hearth was collected for analysis. Several primary roofing elements were collected for tree-ring analysis. Two unfired corrugated vessels and three or four pottery dishes ("pukis") were recovered from the floor and bench of Structure 1316. These suggest that the residents were engaged in pottery manufacture at the time the structure was abandoned. Fill in Structure 1316 consisted of plow zone, a dark organic strata below the plow zone, natural accumulations, and burned roof fall. Structure 1316 appears to have been constructed and occupied in the late A.D. 1200s.

Nonstructure 1317

Nonstructure 1317 consists of a single feature—the base of a complete corrugated vessel excavated into sterile sediment—in the southwest corner of a 2-x-2-m unit (553N 649E). The top of the vessel had been truncated by the plow. Excavation of the remainder of the unit revealed sterile deposits immediately below the plow zone. This may have been a storage feature associated with an extramural work area.

Block 1300 units

Four additional 2-x-2-m units begun last year were completed this season (Table 3). Each revealed sterile deposits immediately below the plow zone.

Block 1400

Excavation within the Block 1400 area was limited to a single structure this season (Figure 8). Additional tree-ring dates were received for samples collected from Structures 1402 and 1408 (Table 1), both burned, masonry-lined kivas completed last season. These new dates support the inference that both structures continued to be used into the A.D. 1250s, and they are consistent with the tree-ring dates previously received for these structures (Duff and Ryan 2000b:Table 2).

Structure 1416

Backhoe Trench 1415 exposed Structure 1416 last season, but no additional work was conducted. In 2000, a unit (Segment 1) was excavated from the base of the trench to the floor of the structure on the west, and a 2-x-2-m unit was placed south of the trench. Segment 1 revealed an earthen and masonry-lined bench face and exposed the floor and a portion of the hearth. Ash from the hearth was collected for analysis. Structure fill consisted of plow zone, a dense trash deposit, and unburned roof fall. The 2-x-2-m unit (586N 660.5E) was excavated from modern ground surface into the trash deposit. This unit was terminated when an adequate sample was collected, although fill within this unit continued. The results of an analysis of the pottery recovered from the midden suggest deposition in the A.D. 1200s, while the architecture of Structure 1416 suggests that it was constructed in the A.D. 1100s.

Block 1500

No excavation was conducted within the Block 1500 area this season, but wood from Structure 1505 produced several tree-ring dates. Structure 1505 is a masonry-lined kiva from which burned roofing elements were sampled in 1999. Tree-ring dates (Table 1) include a cutting date of A.D. 1124 and a noncutting date of A.D. 1155. Pottery from the structure includes Mesa Verde Black-on-white, a type that postdates A.D. 1180 (Wilson and Blinman 1991). The masonry, mode of abandonment, presence of primary roofing elements, and recovery of late pottery types all suggest use of this structure beyond the latest recovered tree-ring date.

Block 1600

The Block 1600 area was first explored in the 2000 field season with three randomly selected 1-x-1-m units (Figure 1). Two of the randomly selected units were adjacent to each other; these were combined and excavated as a single 1-x-2-m unit (755N 342E). The Block 1600 area units were excavated through the plow zone, whereupon sterile sediment was exposed. Artifact density was relatively low.

Block 1700

The Block 1700 area was first explored during the 2000 field season with one backhoe trench and three randomly selected 1-x-1-m units (Figure 9).

Backhoe Trench 1702

Backhoe Trench 1702 was excavated to test a remote-sensing anomaly, but no structure or cultural deposits were found. The trench exposed sterile sediments immediately below the plow zone, and trenching terminated at this depth.

Block 1700 1-x-1-m units

Three randomly selected 1-x-1-m units were excavated this season, each of which exposed sterile sediment immediately below the plow zone. Relatively few artifacts were recovered from these units.

Block 1800

The Block 1800 area was first explored during the 2000 field season with two backhoe trenches, four randomly selected 1-x-1-m units (Figure 9), and six surface collections (Table 2).

Backhoe Trench 1802

Backhoe Trench 1802 was excavated to test a remote-sensing anomaly. The trench was excavated through the plow zone, and sterile deposits were uncovered immediately below the plow zone. Excavations within Backhoe Trench 1802 revealed no structure or cultural deposits.

Backhoe Trench 1803

Backhoe Trench 1803 was also excavated to test a remote-sensing anomaly, but exposed no structure or cultural deposits. Sterile deposits were uncovered immediately below the plow zone.

Block 1800 1-x-1-m units

Four randomly selected 1-x-1-m units were excavated this season. Each uncovered sterile sediment immediately below the plow zone; artifact density was low in all units.

Block 1900

The area of the site not associated with a high-density concentration of artifacts is designated Block 1900, and we continued to conduct surface collections within this portion of the site this season. Surface collections were 3 m in diameter, and their center points are listed in Table 2.

Summary of the Shields Pueblo Investigations

The Crow Canyon Archaeological Center's excavations at Shields Pueblo were completed in 2000, the fourth and final season of field research at the site. All of the excavation units were backfilled. The project resulted in the collection of the types of data needed to address our research questions, and the analysis and reporting of these data will mark a significant contribution to the archaeology of the Mesa Verde region. The surface-collection and random-sampling strategy was applied to each high-density area at the site, and it has provided a systematic sample of artifacts and deposits from across the site. The randomly selected units uncovered several structures and intact deposits that were sampled with additional excavation units. Guided by the results of the remote-sensing surveys, a successful program of structure identification and testing was developed (Duff and Ryan 1998; Duff et al. 1999). This program involved the mechanical testing of remote-sensing anomalies used to define the presence, extent, and fill sequence within structures, portions of which were subsequently excavated by hand. When tested, the majority of the anomalies indicated by the remote-sensing surveys revealed buried prehistoric structures, although some, including those in Blocks 1700 and 1800 which were tested in 2000, did not. Continued analysis of the remote-sensing data, combined with an evaluation of the results of archaeological testing, will help to refine remote-sensing methods and to demonstrate the value and utility of this nondestructive technique for archaeological investigation.

Over the course of four years of fieldwork at Shields Pueblo, approximately 17 burned, masonry-lined kivas, 11 unburned, masonry-lined kivas, 15 earthen-edged or earthen and masonry-lined pit structures, four subterranean rooms, six surface rooms, 68 judgmental 2-x-2-m units, and 90 different 1-x-1-m units were partly or completely excavated by Crow Canyon Archaeological Center researchers and program participants. Several structures had been filled with midden, and some had postabandonment, shrine-like features constructed within them (Ryan 2000). The structures excavated by CMC provided additional data. Information from the Shields investigations has highlighted the variability in historical development of persistent communities in the Mesa Verde region and has demonstrated the potential of heavily disturbed sites to contribute important new data to contemporary research questions.

Shields has provided robust settlement evidence for the Pueblo I through Pueblo III periods. We have documented a few structures that were constructed and occupied during the Pueblo I period (A.D. 700–900), and scattered earlier pottery suggests use of the vicinity during the Basketmaker III period. Shields Pueblo and the McElmo Dome appear to have been abandoned in the A.D. 900s (Varien 1999). The Chaco-era (ca. A.D. 1050–1150) settlement at Shields was much lighter than we expected, as we originally believed Shields to have been the community center during this period. Survey data and the results of previous excavation suggested that the majority of the occupation at Shields dated between A.D. 1050 and 1225. Several pit structures and rooms, primarily in Blocks 100, 200, and 1300, were used in the early A.D. 1100s. The extent of pre-A.D. 1100 occupation at Shields remains uncertain, though pottery assemblages and the architectural characteristics of several structures suggest use during this time. A decline in residential occupation at Shields is apparent in the mid- to late A.D. 1100s. This decline corresponds with a lengthy period of poor environmental conditions (Van West and Dean 2000), and continued analysis of the data will be required to assess the extent of occupation at Shields—if any—in the later A.D. 1100s. Contrary to our expectations, settlement at Shields persisted and grew throughout the A.D. 1200s, with some structures probably occupied until the region was depopulated. Numerous, burned, late-abandoned kivas were examined during the Shields project. The presence of circular masonry features constructed within the depressions created by the collapse of these late-abandoned kivas has not previously been recognized in sites of this period (Duff and Ryan 2000a; Ryan 2000), and the evidence from Shields Pueblo has provided new insights into behaviors associated with the depopulation of the Mesa Verde region in the late A.D. 1200s.

Analysis of all artifacts recovered from Shields is ongoing. Annual descriptive reports, with maps, tree-ring dates, and preliminary interpretations, are already posted on Crow Canyon's Web site. Additional information, updates, and the final report on the work at Shields—including the associated database—will also be made available on Crow Canyon's Web site.