Phase II of the Goodman Point Archaeological Project involves the test excavation of multiple small sites surrounding the large Pueblo Indian (Anasazi) village of Goodman Point Pueblo. This is the interim fieldwork report on the excavations conducted in 2008 by Crow Canyon.
Report of 2008 Goodman Point Community Testing, Montezuma County, Colorado
Grant D. Coffey
In March 2008, the Crow Canyon Archaeological Center launched Phase II of the Goodman Point Archaeological Project: Community Center and Cultural Landscape Study (Kuckelman et al. 2004). Phase I of the study, titled "Goodman Point Pueblo Excavations," was conducted from 2005 through 2007; it involved test excavations at Goodman Point Pueblo (Site 5MT604), the large village that served as the focal point of an extensive community in the A.D. 1200s (Coffey and Kuckelman 2006; Kuckelman and Coffey 2007). Phase II of the project, titled "Goodman Point Community Testing," involves test excavations at multiple smaller sites surrounding the large village, including a variety of habitation sites, ancient roads, and areas that might have been agricultural fields.
All the sites being investigated as part of the Goodman Point Archaeological Project are located in the Hovenweep National Monument, a 142-acre parcel reserved from homesteading in 1889. Crow Canyon’s work in the Goodman Point Unit is being conducted in partnership with the Southeast Utah Group of the National Park Service (SEUG-NPS), which is the managing agency (see ARPA permit number 05-HOVE-01-ext1 for details of the agreement).
The Goodman Point Unit is located in the central Mesa Verde region (Lipe 1995; Varien 2000; Varien and Wilshusen 2002), which in the thirteenth century A.D. was the most densely settled portion of the northern San Juan archaeological region (Figure 1). The unit is also within the Sand Canyon locality, where Crow Canyon has worked for more than 20 years (Lipe ed. 1992; Varien and Wilshusen 2002).
The current phase of the project will test at least 15 habitation and special-use sites within the Goodman Point Unit (Kuckelman et al. 2004). Data gathered through this effort will complement information gained from the first phase of the project, as well as from other archaeological projects carried out in the Sand Canyon locality, and will also result in a more comprehensive occupational history of a large and important ancient community. Testing of selected sites will also provide data that will help us understand the social and environmental adaptations that led to the construction and abandonment of Goodman Point Pueblo.
History of the Goodman Point Unit
The larger "Goodman Point" landform was named after a foreman of the Lacy-Coleman Cattle Company, named Henry Goodman, who drove cattle through the Cortez area in the late 1800s but never settled there. The Goodman Point Unit contains some of the first archaeological resources set aside for protection by the federal government. In 1889, the land tract containing the Goodman Point Unit—Section 4, Township 36 North, Range 17 West—was reserved from homesteading. This action was the result of a recommendation by W. D. Harlan, a U.S. Surveyor General from Denver, Colorado. In 1951, President Harry Truman reduced the size of the protected area to 62 acres within the section and designated this area as part of Hovenweep National Monument. An additional proclamation in 1952 added the additional acreage that composes the present unit, now managed by SEUG-NPS. Because the Goodman Point Unit has been protected since 1889, many sites on the parcel are in pristine condition (Connolly 1992).
Despite its obvious research potential, no systematic testing had been conducted within the Goodman Point Unit before 2005. During the past 50 years, NPS archaeologists have visited the unit to monitor its condition, but research has been largely limited to surface collections at Goodman Point Pueblo, including collections by Pinkley in 1951, by McLellan and Hallisey in 1967, and by an unnamed individual in 1969 (Kuckelman et al. 2004).
Archaeologists from Crow Canyon conducted noncollection pottery tallies at Goodman Point Pueblo in 1986. The results of these tallies, combined with the results of an analysis of sherds gathered during previous NPS collections, indicate that there was a limited occupation of the site during the Pueblo II period and a major occupation during the Pueblo III period (Adler 1986). One of the main goals of our research is to greatly refine the chronology of sites within the unit as a whole and place them in a local and regional context.
As part of a larger survey of the Sand Canyon locality (Adler 1988, 1990, 1992), Crow Canyon archaeologists mapped Goodman Point Pueblo in 1987 using a plane table and alidade. In the same year, they conducted a pedestrian survey of the Goodman Point Unit. This survey focused on residential sites dating from the Pueblo II and III periods, and 17 such sites were recorded as part of this effort. Most of these sites will be tested as part of the current phase of the project.
In 2003, Crow Canyon and the SEUG-NPS conducted a detailed pedestrian survey of the entire 142 acres of the Goodman Point Unit and recorded a total of 42 sites with 56 temporal components (Hovezak et al. 2004). The site density recorded is thus one site per 3.4 acres, or 189 sites per square mile, which is one of the highest recorded densities in the northern San Juan region. The 56 temporal components identified during the survey include four that date from the Basketmaker III period, 15 that are of Pueblo II affiliation, and 23 Pueblo III components (Kuckelman et al. 2004).
From 2005 to 2007, Crow Canyon Archaeological Center researchers compiled an accurate map of Goodman Point Pueblo and conducted test excavations there (Coffey and Kuckelman 2006; Kuckelman and Coffey 2007). In 2008, fieldwork began at seven other sites within the Goodman Point Unit, and that research is the subject of this report.
Research Goals and Strategies
The goals of our research at Goodman Point Pueblo reflect our multifaceted approach to historical, anthropological, and methodological issues, as well as our commitment to American Indian interests. The following provides an overview of some of the broader questions we are addressing; a more detailed discussion of the research goals and objectives can be found in the research design created specifically for this project (Kuckelman et al. 2004).
Historical research goals we are pursuing include assessing the occupational history of the Goodman Point Unit and determining how, when, and why it was depopulated. Anthropological research objectives include examining the settlement ecology of Pueblo farmers in the Mesa Verde region and analyzing how aggregation affects the internal and external organization of communities. Research goals designed to provide information important to American Indian interests include assessing the appropriate methods for studying the relationships between archaeological cultures and modern groups and examining the processes that led to migration from the Mesa Verde region. Lastly, methodological research goals include large-scale goals, such as continuing efforts to produce fine-grained chronologies, and more-specific goals, such as the use of petrographic analysis to produce detailed models of intercommunity exchange.
To achieve these ends, researchers at the Crow Canyon Archaeological Center use field methods and procedures that stress conservation. The specific methods we use in the field can be found in The Crow Canyon Archaeological Center Field Manual (2001). Our practices are guided by the principles of conservation archaeology as outlined by Lipe (1974), which dictate that most of the deposits at a site be left intact for future study. Following from this philosophy, the only artifacts on the modern ground surface that are collected are those found within the actual excavation units; all other artifacts on the modern ground surface are left in place.
The testing strategies used aim to both address the research goals and follow the ethos of conservation archaeology. The size of excavation units used in each context is intended to expose only portions of individual structures or nonstructures (e.g., middens). The careful placement of these excavation units is designed to maximize the information obtained while minimizing the impact to the site.
The nature of the remains being investigated and the type of information desired from specific contexts also guide the size and placement of test units. To facilitate comparisons between different architectural blocks, statistically comparable data from midden contexts are desirable. Therefore, randomly selected 1-x-1-m test units are placed within areas that appear, from evidence visible on the modern ground surface, to possess midden deposits. Additional 1-x-1-m tests units are sometimes placed judgmentally to investigate features or other anomalies present in adjacent units.
Larger units are placed judgmentally in structures and architectural features for the purpose of generating specific data relevant to the research objectives. In kivas, we locate 2-x-2-m units near the southern portion of observable kiva depressions so that we can expose and document architectural features typically found in the southern part of a kiva (e.g., pilasters, southern recess, ventilator tunnel, and deflector) and sample the contents of the hearth. By placing excavation units in these locations, we hope to obtain data relevant to site architectural patterns, subsistence, kiva-related activities, and abandonment treatments.
Kiva depressions are also systematically probed at each site prior to allocating excavation units. Probing is done with a 1-inch soil probe, and individual sediment columns are analyzed and documented by site and depression. Each sample is examined for the presence of charcoal, charred wood, burned sediment, or burned adobe, which might indicate significant structural burning in the kiva. Once probing is completed for the entire site, kiva depressions with the strongest evidence of structural burning are often given priority for testing. Not all kivas with evidence of burning are tested (only 10 kiva 2-x-2-m units for this phase of the project are currently defined in the research design), and, overall, few structures with clear evidence of burning were documented in 2008.
Targeting burned structures for excavation could admittedly skew the data set, but obtaining tree-ring samples will potentially provide a highly precise means of dating particular structures, architectural blocks, and sites. These data will also likely be important for interpreting portions of the site for which no direct chronometric data are available (e.g., unburned kivas and dismantled roomblocks). In addition, even though burned kivas were given priority for testing in 2008, some kivas that produced charcoal and some evidence of burning when probed proved not to be significantly burned upon excavation. These structures will provide data for unburned kivas despite our intentionally targeting burned structures for testing.
Excavation units inside surface structures, along the exterior faces of the north walls of roomblocks, and along the interior faces of architectural enclosing walls (e.g., kiva-enclosing walls) are also placed judgmentally. These 1-x-2-m test units are excavated to expose a variety of architectural elements and cultural deposits. The primary purpose of locating excavation units along the north walls of roomblocks and adjacent to enclosing walls is to document architectural styles and patterns and to reconstruct occupational sequences; test units excavated inside surface structures also yield important information concerning structure use and abandonment practices.
Something that we started doing during Phase II excavations that we did not do during Phase I was to expand some roomblock north-wall units to include a small area inside the structure in an effort to expose interior surfaces or floors. This is done to verify that some poorly preserved wall segments are in fact room walls as opposed to some other type of extramural wall—such as enclosing or retaining walls. The decision to expand the excavation units in this manner is based on architectural content, degree of preservation, and context; we do not expand north-wall units to include room interiors when the portion of wall exposed is sufficient to determine that it is part of a room or other surface structure. The poorer state of preservation of walls at some sites tested in 2008 (compared with walls at Goodman Point Pueblo) is what caused us to modify our excavation strategy.
The great kiva was tested through the excavation of 1-x-3-m units carefully placed to provide information about the architectural style, date, construction, and use of this large and complex structure. So far, four of these units have been placed end-to-end in the southeastern part of the depression; the resulting 12-m-long trench includes a profile of the interior of the kiva and the stratigraphy of the large berm surrounding the structure.
In 2008, we began test excavations in a portion of an ancient road near the Harlan great kiva, called the Goodman Point Belt Loop Road. One 4-x-.5-m test unit and one 2-x-1-m test unit were placed to examine the stratigraphy and artifact assemblage present in the downhill "berm" portion of the road. Though excavation is still in progress, these units will hopefully produce data concerning the construction style and age of this important landscape feature. Crow Canyon staff members Jonathan Till and Jamie Merewether supervised all of the work related to road testing.
Finally, at one small habitation site (Bluebird House), we used a stratified-random-sampling strategy that was first employed by Crow Canyon during the Site Testing Program (part of the larger Sand Canyon Archaeological Project [Varien and Kuckelman 1999]). Specifically, seven sampling strata (roomblock, kiva, courtyard, midden, inner periphery, north outer periphery, and south outer periphery) corresponding to different areas of the site were defined, and 24 randomly selected 1-x-1-m units are in the process of being excavated, with some sampling strata having more excavation units than others. Our goal in employing a random-sampling strategy is to produce data that are directly comparable to the data generated for 13 habitation sites tested during the Site Testing Program (Kuckelman 2004). This strategy should provide data useful in calculating artifact-accumulation rates, which in turn will help us address questions pertaining to the length and continuity of occupation at this site (see Varien and Kuckelman 1999).
All cultural materials and records from the Goodman Point Archaeological Project will be temporarily housed at Crow Canyon until analysis and report preparation are completed. These materials will then be transferred to the Anasazi Heritage Center, a permanent curation facility near Dolores, Colorado.
2008 Field Season
The 2008 field season began on March 31, when field staff began mapping sites and probing kiva depressions for evidence of structural burning. Crow Canyon staff supervised participant excavation at the tested sites from May 5 until October 3. All sites were winterized and closed for the season on December 12.
The first task of the field season was to produce accurate maps of the sites to be tested, which was essential for selecting and setting in excavation units. We mapped each site using a Topcon GT-303 electronic total station surveying instrument and AutoCAD software. The maps produced provide an accurate visual model that we use to delineate and number architectural blocks, associated midden areas, and sampling units. Architectural blocks, as defined at Crow Canyon, are roomblocks and their associated kivas, middens, and extramural areas. Following a convention used for Goodman Point Pueblo, roomblocks at sites with multiple roomblocks were numbered from north to south.
Using numerous instrument-established datums, we mapped kiva depressions, observable wall segments, features, rubble concentrations, and the extent of roomblock rubble for each site. Two mapping transects were completed in the great kiva to provide topographic and elevation data, and the boundary of one site (Bluebird House) was also mapped. Excavation units were also set in using the total station.
Table 1 summarizes, by site and architectural block, the excavation units opened in 2008. The table also specifies which units were completed and which will be continued in 2009. In 2008, 108 test pits were opened and 71 were completed by the end of the field season. Table 1 provides a more detailed breakdown of these units by site and context.
At the close of the field season, units of significant depth that were still in progress were protected with a plywood cover and sealed with plastic sheeting. These measures were taken as safety precautions and to protect the units from damage over the winter. Each completed unit was fully documented and backfilled. A layer of moisture- and vapor-permeable landscaping fabric was placed against all exposed architectural surfaces before backfilling. Great care was taken to place rocks gently against exposed architecture and to backfill each unit with rocks and sediment removed from that unit. The fill was tamped to reduce settling, and the top of the fill was returned as much as possible to the original appearance of the unit at modern ground surface.
In the following sections, the 2008 excavations are summarized by site. The number and location of excavation units set in during the season reflect a focus on the northeast and western parts of the Goodman Point Unit. In 2009 and 2010, we will shift the focus of our investigations to habitation and special-use sites in other areas.
This site is located near the northwest corner of the Goodman Point Unit. Shields Pueblo, investigated by Crow Canyon from 1997 through 2000, is nearby, and it seems likely that Thunder Knoll and parts of Shields Pueblo were occupied simultaneously. Thunder Knoll has at least five discrete roomblocks, six kiva depressions, and four midden areas (Figure 2). Thick sagebrush covers the south-trending ridge on which the site is located, likely obscuring additional subtle surface remains.
We began testing this site in the last two weeks of the field season, and so currently only six excavation units are in progress (one north-wall unit and five midden units in the southern part of the site). None is finished, but from the pottery observed so far, it seems the site has a substantial Pueblo III component, in addition to a slightly earlier, perhaps Pueblo II, component. The southern three roomblocks at the site are marked by very low mounds with small sandstone chunks visible at the surface. It seems likely that the original structures represented by these mounds were either salvaged for building material at some point in the past or were built using a method other than coursed-masonry construction. Fieldwork next year should help clarify when and how major portions of this site were built and how these structures relate to nearby portions of Shields Pueblo.
This relatively large and complex site lies near Goodman Point Pueblo, and it seems likely that portions of Lupine Ridge could have been occupied at the same time as the larger pueblo. Understanding the relationship between these two sites will be critical to reconstructing the late Pueblo occupation of the Goodman Point Unit. Lupine Ridge is an expansive site with approximately 10 to 12 roomblocks, 19 kiva depressions, and at least 10 midden areas spread across a gentle, south-sloping ridge (Figure 3). A dense stand of sagebrush covers the majority of the site.
Initial testing of the site includes 20 excavation units placed in various architectural blocks. In the northern portion of the site, a north-wall unit and two midden units were placed in both Blocks 100 and 300. Though only the midden units have been completed, the surface signatures of the two architectural blocks suggest that the blocks were probably occupied at different times. Block 100 has a substantial rubble mound in the roomblock area, suggesting that this portion of the site was constructed and used later in the occupational sequence than Block 300, which has only a very slight mound and small, sparse rubble in the roomblock area. This pattern may suggest that Block 300 was built first, occupied and then partly dismantled—perhaps with the materials from this block being used to construct later pueblos nearby (e.g., Block 100 or portions of Goodman Point Pueblo).
The block tested most extensively in 2008 is Block 700, located in the east-central portion of the site. This complex area has one roomblock with substantial rubble at the surface and two other low mounds which could be dismantled, possibly earlier, roomblocks. These three areas were grouped together because portions of each architectural component appear to touch, or frame, the area comprising the most visible roomblock rubble mound. In all, as many as eight kiva depressions, three different roomblocks, and three midden areas could be included in this block. One kiva unit, one roomblock north-wall unit, and six midden units were started in Block 700 in 2008.
The kiva unit was placed in a depression that appeared to provide the best chance for recovering burned wood at the site, which could then be dated through tree-ring analysis. Interestingly, no probe samples yielded conclusive evidence of structural burning, although several had small pieces of charcoal and burned adobe. Excavation in this structure, Structure 701, will continue in 2009, but at this point much of the bench and one pilaster have been exposed. Given the absence of burned roof fall on the exposed portion of the bench, it seems unlikely we will recover numerous tree-ring samples from this structure.
The roomblock north-wall unit completed in 2008 provides important information about material reuse and the likelihood of multiple occupations at the site. This unit was placed on a low rise immediately south of the main roomblock rubble mound in an area where an earlier, possibly dismantled, roomblock may have been. This portion of the block was recorded as part of a "multi-terraced complex" during the 2004 survey (Hovezak et al. 2004:72).
Excavation in the north-wall unit revealed the basal courses of a double-stone wall in addition to a formally prepared surface extending from the wall to the south (Figure 4). This surface was, in places, covered with very thin sandstone slabs. Very little rubble was removed from the unit, and the entire depth of excavation from the modern ground surface to the prepared surface was only approximately 28 cm. Three sherds, two pollen samples, and two clay samples were collected from the floor area. The dearth of building stone, the shallowness of the cultural deposits, and the fact that the associated surface was formally prepared suggest the presence of a double-stone masonry room (designated Structure 705) that had been dismantled at some point in the past. This in turn suggests that building materials were reused, with earlier, perhaps late Pueblo II or early Pueblo III, structures being dismantled and the stone being reused to construct later Pueblo III buildings.
The excavation of all six midden units in Block 700 has been completed. Preliminary observation of the pottery assemblage suggests a Pueblo III occupation, with perhaps another poorly defined Pueblo II component. This apparent presence of multiple components will demand much care in the analysis of specific horizontal and vertical provenience relationships if we are to accurately estimate when each part of Block 700 was used.
Two other architectural blocks at the site were tested in 2008: Block 800 and Block 900. The former is located in the southwestern part of the site, and is composed of one roomblock rubble mound, one kiva depression, and one midden area. Only two midden units were excavated in this block to gather preliminary artifact data concerning the period of occupation. Initial impressions of the pottery collected seem to suggest an early Pueblo III occupation.
In Block 900, we excavated in one north-wall unit and three midden units. Excavation in the north-wall unit will continue in 2009, but already substantial midden deposits have been found. Some of this material is likely associated with the occupation of the partly dismantled roomblock to the north (which includes Structure 705). Preliminary observation of the artifact assemblage recovered from the completed midden units suggests a Pueblo III use of the block and perhaps a slightly earlier Pueblo II component.
Some interesting patterns with regard to resource exploitation were noted during the course of the 2008 field season. First, some building materials appear to have been salvaged at the site, a practice probably best reflected in the north-wall unit in Structure 705. This could be important to understanding building stone use and procurement through time. Also of interest is the presence of numerous pit features found in various midden areas throughout the site. In Blocks 300, 700, and 900, eight pit features have been recorded. Although one of these is clearly a thermal feature with fire-reddened margins, the others appear to be primarily borrow pits where native sediment was taken out and later midden material was deposited. The native fill removed from these features could have been used as construction fill in later roomblocks at this site or at nearby Goodman Point Pueblo.
It is likely that more test units will be excavated at Lupine Ridge in the future. Four individual blocks at the site remain untested at this point, and certainly some artifact and architectural information from those blocks would provide a more comprehensive interpretation of the site. This first round of testing was intended to provide baseline data for spatially and temporally diverse areas of the site, data that will hopefully allow us to plan future work.
Pinyon Place lies in the western part of the Goodman Point Unit, near the Harlan great kiva. Three different architectural blocks with five kiva depressions and four midden areas are the main cultural components of the site (Figure 5). Two sandstone concentrations of unknown age and function are also located southwest of Block 200. The site is located in fairly dense pinyon and juniper woodland on a slight, southerly trending ridge.
In all, 16 excavation units were divided among the three architectural blocks in 2008. Blocks 100 and 200 were tested with a roomblock-north-wall and four midden units each, while Block 300 was tested with a north-wall unit, a kiva unit, and five midden units.
Both Block 100 and Block 200 have very low roomblock mounds and very subtle kiva depressions. Small pieces of sparsely scattered sandstone rubble are present in both roomblock locations, and the midden areas are difficult to define spatially due to a paucity of artifacts on the modern ground surface. Only the midden units in these blocks were completed in 2008. From preliminary observations, it seems most of the pottery dates from the Pueblo II period, an interpretation that is also supported by the architectural remains—that is, the presence of small "unit pueblos" (Prudden 1903, 1914) that appear to have been partly salvaged in a manner similar to that posited for Structure 705 at Lupine Ridge.
Evidence from the north-wall unit in Block 200 might reinforce the notion that stone from a masonry roomblock was salvaged at some point in the past. In this unit, we uncovered what appears to be the basal course of a single-stone masonry wall, the vertical-sandstone-slab foundation of that wall, and a prepared floor surface extending from the wall to the south (Figure 6). A wall oriented north-south, south of the north wall, was also uncovered in the unit and displays nearly identical construction. As in Structure 705 at Lupine Ridge, very sparse, small pieces of sandstone rubble were removed from the unit, and the depth from modern ground surface to the prepared floor surface of Structure 206 is only approximately 30 cm.
In the case of Structure 206, the presence of one preserved horizontal course of masonry on top of the vertical sandstone foundation seems to argue for single-stone masonry construction as opposed to the slab-and-jacal construction more characteristic of earlier periods. The lack of rubble, the shallowness of the cultural deposits, the presence of a floor surface, and the intact remains of the structure foundation all argue for at least a partial masonry pueblo, probably dating to the Pueblo II period, that was partly salvaged at some point in the past. Structures of similar age and construction have been recorded in the area (Martin 1938; Robinson and Harrill 1974).
Block 300 displays a surface signature very different from that of Blocks 100 and 200. In Block 300, the roomblock mound is much more pronounced, and the large midden area to the south has a dense concentration of artifacts at the surface. The kiva and north-wall units in this block are still in progress, but all the midden units have been completed. A preliminary review of the pottery collected from these midden contexts suggests at least a Pueblo III occupation of this part of the site—an observation that might be reinforced by the more robust surface signature of the remaining Block 300 architecture. Less Pueblo II pottery noted in the assemblage might also suggest a less-well-defined earlier component. Several excavation units remain to be completed, and laboratory analysis of the artifacts collected is in progress, but currently the northern two blocks appear to date slightly earlier than Block 300 based on the pottery collected so far and the surface signatures of the respective blocks.
Harlan Great Kiva Site
The Harlan great kiva is the largest structure to be tested in 2008. The surface signature of the site is complex and expansive, including one large depression, an extensive berm around that depression, and at least two concentrations of rubble that are likely the remains of surface structures adjacent to the great kiva on the north and south (Figure 7). Four 1-x-3-m units and two north-wall units are being used to test the architecture, and 10 midden units were used to a sample a midden area immediately southwest of the main depression.
At the close of the 2008 field season, seven midden units and one north-wall unit had been completed. Preliminary observation of the artifacts recovered from midden contexts suggests a substantial early Pueblo II use of the site, as well as a less-substantial Pueblo III presence. The presence of numerous Cortez Black-on-white, Mancos Black-on-white, and red ware pottery sherds suggests that the area in which the site is located was used at some point in the late A.D. 900s or early 1000s. This earlier component is somewhat surprising given that analysis of the surface remains recorded during the survey suggested a mid-to-late Pueblo II and Pueblo III use of the site (Hovezak et al. 2004). Determining the relationship between these earlier midden deposits and the great kiva itself (Structure 101) will be an important challenge for future investigations, especially considering the architectural and temporal complexity revealed through excavations this past season.
Though not completed, the test trench formed by excavating four 1-x-3-m units end-to-end in the southeastern portion of the kiva depression has yielded information important to understanding the construction style and period of use. So far, a portion of the great kiva upper-lining wall (the circular masonry wall defining the extent of the main chamber) and an interior masonry feature that is probably a roof-support column and/or a floor vault/foot drum have been exposed in the trench (Figure 8). The masonry used in both constructions incorporates large sandstone blocks that have been flaked but are not as extensively shaped as masonry blocks observed at some late Pueblo III sites, like Goodman Point Pueblo. This might suggest that the latest masonry in Structure 101 was built before the late Pueblo III period, when building styles generally incorporated more extensively shaped blocks. Clearly, more excavation is needed to, first, precisely characterize the masonry elements present in the trench, and then to evaluate those elements for temporal markers.
Numerous small pieces of burned wood recovered from the trench fill so far seem to indicate some degree of structural burning in Structure 101. Though most of the specimens are relatively small, it seems the quantity and overall size of these charcoal pieces is consistent with roofing material that may have burned, or partly burned, and collapsed into the kiva. The presence of an adjacent burned structure immediately south of the trench may complicate interpretations of these remains, however. Hopefully, these samples will help to clarify, more precisely, when the great kiva was built and used.
One of the biggest surprises of the 2008 field season was the recognition that the great kiva itself is smaller than what we had initially estimated on the basis of surface remains. The crest of the earthen berm surrounding the depression, as measured during initial mapping, is 25 to 26 m in diameter; however, excavation revealed that the upper-lining wall defining Structure 101 likely has a diameter of only approximately 13 to 14 m (inference based on the curvature of the exposed portion of the upper lining wall). More excavation is needed to clarify the exact size and morphology of Structure 101, but the observation that the diameter of the main chamber of the kiva is much smaller than the diameter of the earthen berm is important to understanding the overall size, construction, and use of the building.
Perhaps as significant as uncovering interior masonry elements of the kiva was the stratigraphic information provided by testing the large berm that surrounds the depression. Several discrete episodes of deposition are evident in the berm, including an early midden deposit on the prehistoric ground surface, an extensive stratum of redeposited native sediment on top of that, and a thick deposit of midden material that likely represents the final, extensive use of the structure over that (Figure 9). This stratigraphic information suggests at least two periods of intensive use of the site separated by the large-scale excavation and deposition of native sediment from near the structure (perhaps spoil sediment from excavating the great kiva itself?). More-detailed analysis of the stratigraphic profile and associated artifacts will help clarify the time depth represented by the berm, but, overall, a fairly complex picture of occupation and use is already emerging in this area of the site.
Excavations in the north and south ends of the great kiva depression are also helping to define other structures spatially associated with the great kiva main chamber. The 1-x-2-m unit completed near the north end of the depression revealed the presence of a semicoursed masonry wall approximately 1.2 m in length, east to west (Figure 10). This wall is "tied" (created by overlapping stones from one wall into another) to two other wall segments which extend to the south and west. The masonry style of this wall appears similar to that observed for excavated portions of the great kiva interior (flaked but not extensively shaped stones), and it appears the excavation unit exposed only the north wall face of a narrow structure extending to the south. It seems likely, given the curvature of the upper lining wall exposed in the southeastern trench, that this structure lies about 4 m north of the northern extent of Structure 101. The narrow masonry wall exposed in this location protrudes slightly to the north and may be part of a northern antechamber which could have allowed access to the main chamber via a "shaft" or stairway entrance farther to the south. More excavation is needed to confirm this possibility, but antechamber entrances like this have been observed at other great kivas in the region (McLellan 1969).
The 2-x-1-m unit located in the southern portion of the depression has revealed the presence of another structure south of Structure 101. Structure 120 was heavily burned and has plastered walls. Work is ongoing, but at this point there appears to be a purposefully constructed "void" in the south wall of the structure that was later capped with at least one course of masonry. Interpretations of Structure 120 are difficult given the architectural complexity evident in the unit, but it seems the structure could be either an earlier pithouse that was partly excavated through to construct the later great kiva or a plaster-walled, semisubterranean southern antechamber leading into the main part of the great kiva. Hopefully, the numerous tree-ring samples collected from the fill of Structure 120 will help clarify how this structure relates both spatially and temporally to the great kiva. Excavation in an additional unit that was set in immediately north of this excavation unit should also help us interpret the structure.
Bluebird House is a small unit pueblo located near the Harlan Great Kiva site. This site was tested using a stratified-random-sampling strategy similar to that used during the Site Testing Program (Varien and Kuckelman 1999). This strategy calls for mapping the entire site and dividing it into seven different sampling strata, which are then tested randomly, at different rates, to gather data comparable to that collected during the Site Testing Program.
In all, 24 random 1-x-1-m units were divided among the seven different sampling strata at the site, with different proportions of this overall total being allotted to different sampling strata (Figure 11). The proportion of units allocated to each sampling stratum was calculated using methods employed in the Site Testing Program. One judgmental 1-x-1-m unit was also placed in the roomblock stratum to further investigate wall segments and surfaces exposed in adjacent units.
Overall, the extramural cultural deposits at the site were shallow, and some units were nearly devoid of artifacts. Many of the random test units produced small numbers of late pueblo pottery sherds and stone flakes. One unit in the midden sampling stratum, however, produced a bone pendant with turquoise inlay. Very few pieces of turquoise have been found so far during the Goodman Point Project, and determining the source of the turquoise and the approximate age of associated deposits will be important to our assessment of possible trade patterns.
Excavation in the roomblock and kiva areas revealed more stratigraphic and architectural complexity, and as a result, five units in these two sampling strata are not completed. Twenty units placed in other portions of the site are completed at this point, however.
Inferences pertaining to site chronology, the duration of occupation, and artifact accumulation rates will have to await future artifact and statistical analysis, but some preliminary observations seem relevant.
Excavations in three adjacent 1-x-1-m units in the roomblock revealed an architectural pattern apparently similar to that observed in Structure 206 at Pinyon Place. In the westernmost of these units, the foundation of a wall with upright sandstone slabs was exposed. Only a few very small sandstone chunks were removed from the shallow fill above the wall foundation, suggesting that either the masonry portion of the structure had been completely dismantled and removed at some point in the past, or that the roomblock was of some type of poorly preserved, sandstone-and-earth composite construction. Given the preserved portion of the wall at nearby Pinyon Place (which has a very similar foundation construction and one preserved course of masonry overlying that foundation) and the very small amount of rubble removed from the excavation unit, it seems likely that at least a portion of the roomblock had been salvaged at some point in the past.
Excavation in the two units in the kiva sampling stratum is still in progress. Near the end of the field season, a collapsed portion of the southeast upper lining wall was exposed. This suggests that the depression mapped is a collapsed kiva and that we will likely expose a portion of the southeastern part of that structure in the future. Excavations have revealed little evidence of structural burning in the kiva so far.
Monsoon House is a relatively large and architecturally complex site in the western portion of the Goodman Point Unit. The site is composed of four different architectural blocks, with eight or nine kiva depressions, at least seven midden areas, and four rubble scatters (Figure 12). As outlined in the research proposal, six 1-x-2-m units were to be excavated for the purpose of investigating the architectural style of Block 200—a substantial roomblock rubble mound that appears not to have been dismantled in the past and which might have a central plaza area defined by enclosing walls.
In 2008, the site was tested with 22 excavation units spread across the four different architectural blocks at the site. Initial testing included a kiva 1-x-2-m unit (one of the 1-x-2-m architectural-style units prescribed for the site) and three midden units in Block 100; a kiva 2-x-2-m unit, a surface structure unit, an enclosing-wall unit, and eight midden units in Block 200; a surface structure unit in Block 300; and two surface structure units and four midden units in Block 400. Excavation in all the architectural units at the site is still in progress; work in all but one of the midden units has been completed.
Though much work remains to be done at Monsoon House, the 2008 field season produced some interesting preliminary results. First, different architectural components of the site appear to have fundamentally different surface signatures. Block 200, for instance, has a large and highly visible rubble mound, whereas Block 400 has a very low roomblock mound with only scattered small sandstone chunks on the surface. A preliminary survey of the artifacts collected from the midden area of each block also seems to suggest slightly different occupational histories, with more Pueblo II pottery in the Block 400 assemblage compared with the assemblage from Block 200, which has more Pueblo III pottery. The remaining portions of the walls exposed in Block 200 so far also have more substantial masonry intact, perhaps suggesting later construction. Further lab analysis and completion of the remaining architectural excavation units should help to clarify if indeed an earlier component is present in the southern area compared with the more northerly blocks.
Taken as a whole, the results of our 2008 excavations suggest the presence of at least a couple of temporally discrete occupations of the site spanning at least several decades. Future excavations at the site will likely entail the testing of two additional midden areas at the site and placing additional 2-x-1-m units in Block 200 to explore this complex and interesting architectural area.
Fieldwork in 2008 included the initiation of two important additional field studies. These studies, outlined in the following sections, should produce data that will complement and enhance information gathered through test excavations. Hopefully, these studies will be the first of many that examine archaeological, environmental, and ecological variables relevant to understanding the physical and social contexts that impacted ancient communities in the Goodman Point Unit.
In August 2008, exploratory remote sensing was conducted at several sites within the Goodman Point Unit. William Volf, of the National Resources Conservation Service (NRCS), volunteered his time, equipment, and expertise to conduct several remote-sensing transects using two highly sophisticated instruments: an electricity-resistance meter and a device called a magnetometer. These high-tech instruments locate subsurface archaeological remains by shooting a low-voltage electrical current through the ground (the electricity-resistance meter) and by recognizing magnetic anomalies that could indicate the presence of buried structures and features (the magnetometer).
The purpose of this pilot effort was twofold. First, we wanted to look at the archaeological and practical applicability of using the instruments on the densely vegetated terrain of the Goodman Point Unit, and second, we wanted to use nearby excavation units as a reference for interpreting data gathered by each instrument. This effort appears to have been at least partly successful at two sites (Harlan Great Kiva and the Goodman Point Belt Loop Road), while at another site (Monsoon House) the data are less clear. Also, the RM-15 Resistance Meter appeared to produce the most interpretable results (at this point anyway), while the FM-256 Fluxgate Gradiometer produced data that are more difficult to decipher. Therefore, only the resistivity data are discussed here.
We used the resistance meter at the Harlan great kiva, completing a transect immediately north of the 12-m trench we are currently excavating. The exposed portion of the upper lining wall, revealed through excavation, appears to correlate well with the highest resistance recorded on the transect. If the curvature of the exposed wall is taken to the opposite side of the structure, it appears another highly resistant area there may represent the buried upper lining wall in that location (Figure 13).
We also conducted a resistance-meter transect on a portion of the "belt-loop" road near the Harlan great kiva. The data appear to be less readily interpretable in this case, but the device again shows an area of high resistance in the space roughly defined by the downhill "berm" of the road. Though preliminary in nature, these data may indicate an area of disturbance consistent with the "piling" of sediment created by excavating and flattening part of the ancient road upslope.
Hopefully more remote-sensing studies will be completed in future years of the project. The results so far indicate that the resistance meter in particular might be useful for selecting and cross-referencing new excavation units. Systematically gridding and mapping portions of low rubble mounds and other amorphous features using remote-sensing instrumentation might be particularly useful. None of the remote-sensing work completed in 2008 would have been possible without the generous support of the NRCS and William Volf.
In 2008, a series of temperature monitors were placed in various locations within the Goodman Point Unit. These electronic monitors record very precise temperature data and can operate unaided for months at a time. Monitors were placed on upland south slopes, on upland north slopes, on drainage hillsides, and in drainage bottoms to collect data fairly representative of the topographic variability present in the Goodman Point Unit.
The purpose of these monitors is to collect long-term temperature data for various parts of the Goodman Point Unit that could have once been Pueblo agricultural fields (the goal is to have the monitors in place for at least three years). The data collected will be studied for patterns of temperature variation which might point to certain areas of the Goodman Point Unit being more or less favorable for agriculture. Length of frost-free periods and number of corn growing-degree days will likely be studied for each location in order to assess the agricultural potential of various locales within the unit. These temperature data will hopefully be combined with other environmental and ecological data to study potential agricultural productivity in areas where ancient farmers may have grown crops. Crow Canyon staff member Ben Bellorado is overseeing all aspect of this study.
Summary and Interpretations
The first year of Phase II research at the Goodman Point Unit was very successful in terms of obtaining data relevant to our research goals. The construction of high-resolution maps of the tested sites, the documentation of individual excavation units, and the recording of artifact data and stratigraphic information provided a solid foundation for interpreting the occupational history and chronology of sites tested in 2008.
Some important observations relating to the research goals were also made during the 2008 field season. Among these were the recognition of the smaller horizontal extent and chronological complexity of the Harlan great kiva. Original estimates suggested a main chamber with a diameter of approximately 22 to 24 m based on depression morphology and berm configuration (Hovezak et al. 2004:101), but test excavations that exposed some interior masonry indicated a much smaller diameter for the kiva—approximately 13 to 14 m. In tandem with this observation, the large earthen berm surrounding the great kiva appears to be much larger and more substantial than originally thought. This might suggest that the constructed berm served an important practical, aesthetic, or symbolic role in the construction of the great kiva and, by extension, to the community which created it.
Tree-ring samples collected from inside the great kiva itself, as well as from nearby structures, should help clarify when the structure was built and used and how it relates to other nearby structures. Completing the 1-x-3-m units in the great kiva interior will provide an uninterrupted 12-m profile that should allow us to make solid inferences concerning the construction style and use life of the structure. Additional units placed in and around the great kiva depression should also produce stratigraphic, architectural, and assemblage data that will be central to forming a more comprehensive view of this impressive building.
The presence of an apparent early Pueblo II component at the Harlan Great Kiva site also presents some interesting questions for future research. Pottery cross-dating during the survey suggested that the great kiva was used during the late Pueblo II period (Hovezak et al. 2004:101), and while this may still prove to be true, excavations in various parts of the great kiva area revealed a substantial quantity of earlier Pueblo II pottery (e.g., Cortez Black-on-white and Mancos Gray). This observation leads to questions about when and how the great kiva area was first used, and what influences this early use of the site may have had on later developments.
Though early Pueblo II great kivas are few in number in the Mesa Verde region, one excavated in Morfield Canyon (MV1067), within Mesa Verde National Park, may provide an interesting comparison for what has been observed so far at the Harlan great kiva. In the Morfield Canyon example, Cortez and Mancos black-on-white are the dominant decorated pottery types found in both the upper and lower fill of the structure, with far fewer sherds of McElmo Black-on-white and Mesa Verde White Ware present. That great kiva also has a masonry upper lining wall, antechamber entrances, and interior masonry features that could be called floor vaults or foot drums. The diameter of the Morfield structure ranges from approximately 13.7 m to 15.24 m, despite what appear to be at least two episodes of remodeling (McLellan 1969). The Harlan great kiva appears to display many of these same characteristics, perhaps suggesting a roughly contemporaneous use of the structure. In any case, the Morfield great kiva appears to be a useful point of reference for future work at the Harlan Great Kiva site.
Potentially important observations regarding material reuse were also made in 2008. Excavation in several roomblock north-wall and surface structure units revealed very little building stone, shallow cultural deposits, and—in some cases—formal, prepared surfaces associated with wall foundations. While it is possible that some roomblocks tested were built using methods other than masonry construction, the overall dearth of construction material of any kind recovered from test excavations, the presence of stone wall foundations, usually the absence of roof-fall material in unit fill, and the associated formal surfaces seem to suggest extensive reuse of building materials through time. Many of the excavation units still need to be completed, but it seems this pattern has been identified in roomblocks in at least three sites. Also, roomblocks within the same site, or at sites very nearby, display much more pronounced rubble mounds, which were apparently not salvaged extensively in the past. This could suggest that extensive salvaging did take place at very precise times in the ancient past—likely occurring on a large scale during the Pueblo III period.
Several of the midden areas tested also had fairly shallow deposits in general. These relatively shallow midden deposits, and the presence of numerous midden areas at most sites, may suggest frequent relocation of individual roomblocks or pueblos through time within the larger site area. Possible evidence of extensive reuse of building materials may reinforce this notion. Several of the sites tested also appear to have both early and later Pueblo occupations, perhaps suggesting that while site locales were frequently reused through time, individual roomblocks were occupied for relatively short periods. Better defining possible "gaps" in the occupational history of these sites will be important to presenting better individual histories for the sites tested and for the Goodman Point Unit as a whole.
Another interesting pattern concerning roomblock characteristics was noted in 2008. At two of the sites (Lupine Ridge and Pinyon Place), tested roomblocks that appear to be at least partly dismantled outnumber intact roomblocks by at least two-to-one. Although not all of the roomblocks have been tested, by using the surface signatures of those that have been tested, along with the excavation data gathered so far, it seems that earlier roomblocks at some sites are smaller and more numerous than "intact" roomblocks, which seem to be larger and fewer in number. Future testing and analysis will hopefully help us determine if this pattern is valid, and if so, what this alteration in settlement patterning might mean for interpreting community formation, aggregation, and adaptation for the subsequent Pueblo sequence in the Goodman Point Unit.
In 2008, a large and diverse segment of the interested public benefitted from Crow Canyon's research during the first year of Phase II of the Goodman Point Archaeological Project. The excavation portion of the project involved the participation of 494 people enrolled in Crow Canyon research and education programs, including schoolchildren, teens, and adults. Numerous formal tours given as part of our "day" programs, as part of our nonexcavation school curriculum, or as part of other Crow Canyon–sponsored activities resulted in at least 800 other individuals learning about our research at the Goodman Point Unit.
These figures reflect Crow Canyon's commitment to involving diverse segments of the public in our research and also demonstrate the level of public interest in the ancient past of the Mesa Verde region. This research and education effort would not have been possible without the cooperative partnership between Crow Canyon and the NPS.
American Indian Involvement
American Indian consultation is an important part of the research process at the Crow Canyon Archaeological Center. On April 26, 2008, Acoma Pueblo elder and Crow Canyon Native American Advisory Group member Ernest M. Vallo, Sr., conducted a blessing ceremony at the Harlan great kiva before Phase II excavations began. The ceremony was attended by consultants from several regional tribes and many members of Crow Canyon's Board of Trustees.
As part of a continuing cooperative effort between Crow Canyon and American Indian consultants, the Pueblo Farming Project continued in 2008. Meetings held in May and October focused on planting and harvesting crops, respectively, in various experimental plots on the Crow Canyon campus. This cooperative effort was initially outlined in the research proposal for the Goodman Point Unit (Kuckelman et al. 2004), and information gathered will hopefully result in both an educational curriculum for American Indian students and scientific data that can be used to model agricultural productivity in the area. Crow Canyon staff members Ben Bellorado, Paul Ermigiotti, and Mark Varien have been working closely with representatives of several tribes on this project. The blessing ceremony and the continuing consultation and involvement of American Indian groups throughout the Goodman Point Archaeological Project demonstrate the Center's dual commitment to archaeological field research and working with American Indians in the study of the past.
Plans for the 2009 Field Season
In 2009, we will continue to focus on the testing of habitation sites surrounding Goodman Point Pueblo. This effort will include finishing excavations in units begun last year and testing additional portions of Lupine Ridge and Monsoon House; in addition, we will open new units at the Harlan Great Kiva site. As we complete those excavations, we will begin shifting our focus to the northern part of the Goodman Point Unit, where we will initiate test excavations at several new habitation sites. Finally, we hope to begin testing possible agricultural fields and checkdams in the Goodman Point Unit, as well as continue the temperature and remote-sensing studies started in 2008.
The Crow Canyon Archaeological Center's 2008 field and laboratory program related to the Goodman Point Archaeological Project was funded in part by State Historical Fund grants from the Colorado Historical Society.
Research Field Personnel, 2008 Field Season
Grant Coffey, supervisory archaeologist
Steve Copeland, research archaeologist
Susan Ryan, research archaeologist
Jonathan Walker, seasonal archaeologist
Charlie Reed, research intern
Jacob Sedig, research intern
Harrison Ignacio, research intern
Perri Gerard-Little, research intern
Brigitte Wray-Miller, research intern
Ewa Michoń, research trainee
Education Staff Personnel, 2008 Field Season
Others Who Worked or Volunteered, 2008 Field Season
Chris Goetze (NPS)
Corky Hays (NPS)
Erin Lewis (NPS)
Laura Martin (NPS)
Chris Nickel (NPS)
William Volf (NRCS)
Adler, M. A.
1986 Report on a Non-collection Reconnaissance of Goodman Point Ruin. Manuscript on file, Crow Canyon Archaeological Center, Cortez, Colorado.
1988 Archaeological Survey and Testing in the Sand Canyon Pueblo/Goodman Point Ruin Locality, Montezuma County, Colorado, 1987 Field Season. Crow Canyon Archaeological Center, Cortez, Colorado. Report submitted to the Bureau of Land Management, San Juan Resource Area Office, Durango, Colorado.
1990 Communities of Soil and Stone: An Archaeological Investigation of Population Aggregation Among the Mesa Verde Region Anasazi, A.D. 900–1300. Ph.D. dissertation, University of Michigan. University Microfilms, Ann Arbor.
1992 The Upland Survey. In The Sand Canyon Archaeological Project: A Progress Report, edited by W. D. Lipe, pp. 11–23. Occasional Papers, no. 2. Crow Canyon Archaeological Center, Cortez, Colorado.
Coffey, G. D., and K. A. Kuckelman
2006 Report of 2005 Research at Goodman Point Pueblo (Site 5MT604), Montezuma County, Colorado. Crow Canyon Archaeological Center, Cortez, Colorado. Report submitted to the Southeast Utah Group of the National Park Service, Moab, Utah.
Connolly, M. R.
1992 The Goodman Point Historic Land-Use Study. In The Sand Canyon Archaeological Project: A Progress Report, edited by W. D. Lipe, pp. 33–44. Occasional Papers, no. 2. Crow Canyon Archaeological Center, Cortez, Colorado.
Crow Canyon Archaeological Center
2001 The Crow Canyon Archaeological Center Field Manual [HTML Title]. Available: https://www.crowcanyon.org/fieldmanual.
Hovezak, T. D., L. M. Sesler, M. D. Varien, and C. Goetze
2004 An Archaeological Survey of the Goodman Point Unit of Hovenweep National Monument, Montezuema County, Colorado. Crow Canyon Archaeological Center, Cortez, Colorado. Report submitted to the Colorado Historical Society, Denver, and the Southeast Utah Group of the National Park Service, Moab, Utah.
Kuckelman, K. A., M. D. Varien, S. G. Ortman, and J. D. Till
2004 A Proposal to Conduct Archaeological Testing at the Goodman Point Ruins Group Unit of Hovenweep National Monument. Manuscript on file, Crow Canyon Archaeological Center, Cortez, Colorado.
Kuckelman, K. A., and G. D. Coffey
2007 Report of 2006 Research at Goodman Point Pueblo (Site 5MT604), Montezuma County, Colorado [HTML Title]. Available: https://www.crowcanyon.org/goodmanpoint2006.
Lipe, W. D.
1974 A Conservation Model for American Archeology. The Kiva 39:213–245.
1995 The Depopulation of the Northern San Juan: Conditions in the Turbulent 1200s. Journal of Anthropological Archeology 14:143–169.
Lipe, W. D. (editor)
1992 The Sand Canyon Archaeological Project: A Progress Report. Occasional Papers, no. 2. Crow Canyon Archaeological Center, Cortez, Colorado.
Martin, P. S.
1938 Archaeological Work in the Ackmen-Lowry Area, Southwestern Colorado, 1937. Anthropological Series, vol. 23, no. 2. Field Museum of Natural History, Chicago.
McLellan, G. E.
1969 The Origins, Development, and Typology of Anasazi Kivas and Great Kivas. Ph.D. dissertation, University of Michigan. University Microfilms, Ann Arbor.
Prudden, T. M.
1903 The Prehistoric Ruins of the San Juan Watershed in Utah, Arizona, Colorado, and New Mexico. American Anthropologist 5:224–288.
1914 The Circular Kivas of Small Ruins in the San Juan Watershed. American Anthropologist 16:33–58.
Robinson, W. J., and B. G. Harrill
1974 Tree-Ring Dates from Colorado V: Mesa Verde Area. Laboratory of Tree-Ring Research, Tucson.
Varien, M. D.
2000 Introduction. Kiva 66:1, pp. 5–18.
Varien, M. D., and K. A. Kuckelman
1999 Introduction. In The Sand Canyon Archaeological Project: Site Testing, edited by M. D. Varien, Chapter 1. CD-ROM, Vers. 1.0. Crow Canyon Archaeological Center, Cortez, Colorado.
Varien, M. D., and R. H. Wilshusen
2002 A Partnership for Understanding the Past: Crow Canyon Research in the Central Mesa Verde Region. In Seeking the Center Place: Archeology and Ancient Communities in the Mesa Verde Region, edited by M. D. Varien and R. H. Wilshusen, pp.3–23. The University of Utah Press, Salt Lake City.
REMINDER: Archaeological resources are protected by federal laws, and archaeological research is guided by a set of professional ethics. See Archaeological Ethics and Law.
© Copyright 2009 by Crow Canyon Archaeological Center. All rights reserved.