The Archaeology of Woods Canyon Pueblo

Chronology

by Melissa J. Churchill and Scott G. Ortman

1
In this chapter, we attempt to reconstruct the occupational history of Woods Canyon Pueblo through an analysis of tree-ring data, archaeomagnetic dating results, structure locations and architectural characteristics, structure abandonment mode, stratigraphic evidence, and pottery data. We examine these various lines of evidence by grouping survey and excavation data from the site into seven areas within four sections of the pueblo. In "Architecture and Site Layout," Churchill uses natural topographic boundaries to divide Woods Canyon Pueblo into four sections: the canyon bottom, east talus slope, upper west side, and canyon rim. These four sections have been further subdivided into numbered areas (Areas 1–7) in an attempt to group together excavation units that reflect the spatial associations among specific structures and nonstructures (see Database Map 334).

2
The canyon bottom is subdivided into two areas. Area 1 includes test units that fell inside Structure 1-S and its associated midden (Nonstructure 4-N); test units in Nonstructure 1-N, a large, flat, open area, part of which is likely associated with an unexcavated kiva to the north; and test units in Structure 9-S, which was discovered in the course of testing Nonstructure 1-N. The information presented in this chapter for Area 1 thus derives from our sampling of all three kiva suites in the west half of the canyon bottom. Area 2 includes test units that fell inside Structures 2-S and 3-S, two of the eight to 10 kivas clustered in the east half of the canyon bottom, and test units in Nonstructure 5-N, which encompasses the middens apparent on the modern ground surface in this area. Information for Area 2 thus derives from sampling the kiva suites in the east half of the canyon bottom.

3
Tested structures and nonstructures on the east talus slope have also been grouped into two areas. Area 3 includes the test unit that fell inside Structure 4-S, units placed in its associated midden (Nonstructure 8-N), and material from the original 1-x-2-m test unit that was intended to fall inside Structure 4-S, but ended up falling just outside and to the west of the structure (the cultural deposits in this unit were designated Nonstructure 12-N). Accordingly, information for Area 3 derives from sampling the specific kiva suite focused on Structure 4-S. Area 4 includes test units inside Structure 5-S and its associated midden (Nonstructure 6-N); therefore, information for this area derives from sampling the specific kiva suite focused on Structure 5-S.

4
Tested cultural features on the upper west side have been grouped into two areas. Area 5 includes test units placed inside Structure 7-S and its associated midden (Nonstructure 3-N). Because Nonstructure 3-N is located to the south and downslope of as many as two additional, but untested, kivas, it is likely that our excavations sampled deposits associated with all three kiva suites, not just the one focused on Structure 7-S. Area 5 also includes a 1-x-2-m unit that was intended to fall within yet another kiva, the presence of which had been suggested by surface evidence; however, when excavation failed to reveal a structure, the cultural deposits within this unit were designated Nonstructure 10-N. Because of the likely association of Nonstructure 3-N with as many as three kiva suites, the information presented in this chapter for Area 5 derives from these multiple suites in the northwest corner of the site. Area 6 includes test units placed inside Structure 8-S and its associated midden (Nonstructure 7-N). Again, because this midden lies to the south and downslope of as many as two additional kivas, and therefore is likely to be associated with them as well, information for Area 6 derives from our sampling of these multiple kiva suites along the base of the cliff and immediately west of the drainage that bisects the site.

5
Finally, Area 7 comprises all excavation units in the canyon rim section, including those in the rim complex proper. Although testing in Area 7 included the excavation of units located across checkdams and possible agricultural terraces in Nonstructure 9-N (located west of the enclosing wall that defines the rim complex), very little material of chronological significance was found in these units, so it is appropriate to associate information for Area 7 with the rim complex itself.

6
When Crow Canyon began its research at Woods Canyon Pueblo, several lines of evidence suggested that the site dated primarily from the middle and late A.D. 1200s (Lipe 1995*2; Varien et al. 1996*1; Wilshusen et al. 1997*1). Our excavations have revealed that this interpretation is inaccurate. Although the excavation data do substantiate that Woods Canyon Pueblo was a large village during the late Pueblo III period (A.D. 1225–1280), several lines of evidence suggest that a portion of the site was also occupied earlier in the Pueblo III period (A.D. 1140–1225) and may have fallen out of use a few decades before the end of occupation at the site (Ortman et al. 2000*1). Thus, we now believe that Woods Canyon Pueblo was occupied throughout the Pueblo III period, from approximately A.D. 1140 until the final emigrations of Pueblo people from the Mesa Verde region in the late 1200s. The following sections present the data upon which these conclusions are based.

Tree-Ring Dating

7
We used tree-ring dates to establish when particular structures or architectural complexes were built at the pueblo. These dates, more than any other type of information, support the interpretation that there was a late Pueblo III occupation at the site. During surface survey and excavation, we collected 195 tree-ring samples; of these, 93 yielded dates (Figure 1). Most of the dated samples were construction timbers: 72 were collected from the modern ground surface in and downslope of the rim complex, and 18 derive from test units in kivas or from in situ wood inside structures protected by the cliff. Only three of the dated samples were collected from nonstructural contexts.

8
On the basis of earlier work by Dean (1978*2), Ahlstrom (1985*1) developed a number of principles for interpreting tree-ring dates associated with construction timbers. Two of these principles are particularly relevant to the interpretation of the Woods Canyon tree-ring data. First, it is assumed that timbers were harvested specifically for construction, and therefore each timber would have been used within a year or two of the date of its outermost growth ring (Ahlstrom 1985*1:58; Dean 1978*2:148). Second, in the absence of date clusters (as is the case at Woods Canyon Pueblo), the latest date associated with a structure provides the best estimate of when construction occurred (Ahlstrom 1985*1:58–59). At a minimum, it can be assumed that a structure was in use at least as recently as its latest associated date (Dean 1978*2:148). Additionally, if the latest date for a structure is a noncutting date—that is, a date that predates the year that the tree died, usually because the outer rings on the sample are damaged or missing—it is safe to assume that the structure was constructed or remodeled some time after that date. We relied heavily on this assumption in our analysis of the Woods Canyon Pueblo tree-ring dates because our work resulted in only two cutting dates being obtained for the entire site.

9
A noncutting date for a sample collected from an in situ roof beam suggests that at least one building at Woods Canyon Pueblo (Structure 10-S and its associated upper story, Structure 20-S) was constructed after A.D. 1276. This is the latest date for the site, and it is believed to be associated with the initial construction of the building rather than with a remodeling episode because it was obtained for a primary beam that extends through the north wall. It would have been difficult, if not impossible, to add this beam during remodeling without destroying the wall.

10
The overwhelming majority (84 percent) of the tree-ring samples that yielded dates derive from structures in the rim complex. The rim complex includes the buildings and space enclosed by a wall on the canyon rim, as well as the buildings at the base of the cliff directly below the rim. Most of these dates were yielded by samples collected from the modern ground surface during initial mapping in 1993 (Lipe 1995*2:2); the samples were from beams found close to, but outside, surface rooms. Even though the samples were extremely weathered and produced only noncutting dates, the Laboratory of Tree-Ring Research was able to provide a date range of A.D. 1240–1260+. This inference was based on estimates of how many years were represented by the missing sapwood rings:

Since sapwood rings may easily involve the outer 50 to 100 rings in juniper, this means that the "vv" [noncutting] dates are, in general 50 to 100 years earlier than the true [cutting] date. The few samples that have a few sapwood rings all date post–[A.D.] 1200, even though they, too, are weathered. Thus from a pure dating point of view, it is likely that all the samples should fall in the [A.D.] 1240 to 1260+ time frame [Robinson 1993*1].

11
Tree-ring samples were also obtained from several individual structures in the rim complex (Table 1 and Database Map 332). Core samples were collected from 13 beams, many of which were found in situ in standing buildings. On the basis of noncutting dates, we believe that two ground-story rooms (Structures 10-S and 11-S) and their associated upper-story rooms (Structures 20-S and 21-S) were constructed after A.D. 1276 and A.D. 1271, respectively. Because these dates were obtained for in situ roof beams embedded in the room walls, we believe that they are associated with the initial construction of the buildings rather than with a remodeling episode.

12
A weathered, in situ beam in a kiva (Structure 13-S), located just east of these rooms, produced a noncutting date of A.D. 1212. However, the upper lining wall of this kiva abuts the east wall of Structures 11-S and 21-S, which must have been built after A.D. 1271, given the noncutting date obtained for the in situ beam mentioned above. It is therefore most probable that this kiva (Structure 13-S) was also built after A.D. 1271. The timber that yielded the A.D. 1212 noncutting date was heavily eroded, and it also could have been salvaged from an earlier structure. Timbers found in the wall and roof fall of Structures 6-S and 23-S also yielded noncutting dates of A.D. 1204 and 1268, respectively. In summary, the tree-ring data and inferred construction sequences clearly indicate that numerous structures in the rim complex were constructed in the mid- to late 1200s.

13
Tree-ring samples were not as useful in dating other parts of the site (Database Map 331). Fewer beams were found in kivas than was expected, probably because the roofs were dismantled after the structures were abandoned, or perhaps because some of the roofs were not completely burned (unburned wood decomposes more quickly and completely than does burned wood). On the east side of the pueblo (in Area 4), the latest noncutting date for a tree-ring sample from the roof-fall deposits of a kiva (Structure 5-S) is A.D. 1143. On the upper west side of the pueblo (in Area 5), a single cutting date for Structure 7-S, a kiva, is used to infer construction or remodeling in A.D. 1257.

14
A tree-ring sample collected from Nonstructure 3.3-N, a midden in Area 5, yielded a noncutting date of A.D. 1231. The only tree-ring date for the canyon bottom is a noncutting date of A.D. 1157, from a sample collected in Nonstructure 1.18-N, a midden in Area 1. Because fuelwood could have been dead for many years before being collected and burned, these dates from midden samples may predate activities in Areas 1 and 5 by many years. Nevertheless, they document human activity during the Pueblo III period in these areas. So in addition to substantiating late Pueblo III construction in Area 7, tree-ring data also indicate that Areas 1, 4, and 5 were scenes of human activity sometime during the Pueblo III period, between A.D. 1140 and 1280.

Archaeomagnetic Dating

15
Archaeomagnetic samples collected from two hearths in Structure 2-S, a kiva located in Area 2 (in the canyon bottom), were analyzed by the Archaeometric Laboratory at Colorado State University; there are no tree-ring dates for this structure. The date ranges for the earlier hearth (Feature 8) are A.D. 1000–1025 and A.D. 1200–1325. Because Pueblo III pottery was found on the floor of the kiva, the latter date range seems the most plausible; if it is accurate, the hearth was used after A.D. 1200. The later hearth (Feature 3)—built inside the earlier hearth—provided date ranges of A.D. 1325–1475 and A.D. 1675–1750. Both ranges are later than the ancestral Pueblo occupation of the Mesa Verde region. Whether these date ranges indicate a compromised sample, inaccuracy in the archaeomagnetic dating method, or post-Puebloan activity at the site is unknown; however, because no other evidence suggestive of post-Puebloan activity was discovered in the course of our research, the last explanation is the least likely of the three.

Structure Location

16
Varien (1999*5) used survey and excavation data from the Sand Canyon locality to identify trends in site location during the Pueblo III period. He found that most mesa-top sites had fallen out of use by the mid–A.D. 1200s and that most sites occupied during the mid- to late 1200s were located on canyon rims, at the bases of cliffs, or on the talus slopes of canyons. Areas 1 and 2 of Woods Canyon Pueblo are located in the canyon bottom, a setting that was not investigated intensively during the Sand Canyon Project. However, Areas 3–7 contain structures located on the canyon rim, the base of the cliff, and the talus slope of the canyon. These were all favored settlement locations during the late Pueblo III period in the Sand Canyon locality.

Architectural Styles

17
Most of the structure walls that we documented at Woods Canyon Pueblo were constructed entirely of stone masonry. Room walls and the lining walls of kivas are almost always semicoursed and made of rectangular sandstone blocks that were pecked on at least one face. This architectural style is typical of structures dating from the Pueblo III period (A.D. 1140–1280) in the Mesa Verde region (Lipe and Varien 1999*1). In addition, most of the room and tower walls still standing above the modern ground surface are double-stone wide and have rubble cores. Double-stone walls first appeared in the Mesa Verde region in the late 1000s and early 1100s and were commonplace in structures built during the 1100s and 1200s (Varien 1999*3).

18
Two of the tested kivas in the canyon bottom were only partly lined with stone masonry. These structures (Structures 1-S and 9-S, in Area 1) have masonry bench faces but earthen upper lining walls. Smith (1998*1) recently analyzed the evolution of kiva architecture in the Mesa Verde region and found that more than 50 percent of kivas constructed between A.D. 1150 and 1200 were partly earthen walled and partly lined with stone masonry. The lower lining walls, benches, and pilasters of these structures were typically made of stone, and the upper lining walls were earthen. In contrast, more than 90 percent of post–A.D. 1200 kivas were completely lined with stone masonry. Thus, the part-masonry, part-earthen building style of the tested kivas in Area 1 at Woods Canyon Pueblo suggests that these kivas were constructed before A.D. 1200. Earthen upper lining walls were not found in kivas we tested in other areas of the village.

19
A discrete cluster of buildings that were likely constructed early in the occupation of Woods Canyon Pueblo was also identified during surface survey of the canyon bottom. These buildings are located at the southernmost end of the site, approximately 20 to 30 m north of the drainage in the bottom of the canyon (Kelley 1996*1:67). They were recorded during the 1993 mapping project as Features 151–155 (see Appendix A) but were not tested. Surface remains suggest that this cluster includes two towers, several kivas, and mounds of earth and stone that probably are the remnants of surface structures. The mounds are unlike other rubble mounds at the site in that they contain more earth than stone. For the most part, the stones are small, few in number, and widely scattered. It is possible that the buildings in this area were constructed of pecked-block masonry that was dismantled later during the Pueblo occupation; it is also possible that they were made of a combination of earth and stone, with the former being the dominant construction material. Either scenario would suggest that these buildings were made relatively early in the history of the village.

Structure Abandonment Data

20
On the basis of data from the Dolores Archaeological Program and the Sand Canyon Project, Lightfoot (1993*1), Schlanger and Wilshusen (1993*1), and Varien (1999*1) have developed middle-range theory that relates roof treatments and floor assemblages to household mobility. When a household moved to a new residence a short distance away and did not anticipate returning to the old house, the people tended to salvage the roof timbers and take usable household items to the new location, resulting in few artifacts or roof timbers being left behind in the abandoned structures. In contrast, when people moved long distances and did not anticipate returning, they often burned the roofs and left behind many household items because of limitations on what they could carry over long distances. Homes abandoned during the final regional emigrations of the late A.D. 1200s were left by people intending to make long-distance moves and not return. Hence, roof treatment and floor-assemblage data can assist us in identifying areas of Woods Canyon Pueblo that date to the final decades of Puebloan occupation in the Mesa Verde region.

21
Varien (1999*1) classified the roof-fall deposits of Pueblo III kivas in the Sand Canyon locality into four interpretive types: (1) burned intact, (2) burned and salvaged, (3) unburned salvaged, and (4) unburned unsalvaged. He found that "burned intact" and "burned and salvaged" roofs occur primarily in structures that were in use at the time of the final emigrations of Pueblo people from the Mesa Verde region. In contrast, the roofs of kivas dating to earlier periods most often were unburned and salvaged, presumably so that construction timbers could be reused in new structures being built nearby.

22
The roof treatment of Structures 6-S and 13-S in the rim complex at Woods Canyon Pueblo could not be determined. However, the tested kivas in Areas 3, 4, 5, and 6 contain burned and salvaged roofs (see "Abandonment and Emigration"). In addition, the roof of Structure 5-S (in Area 4) had collapsed directly on top of a mass interment containing remains of at least 10 individuals. In "Human Skeletal Remains," Bradley argues that this building was a sealed sepulcher or tomb for some period after the human remains were interred. An alternative interpretation is that the roof was dismantled and collapsed on top of the remains immediately after interment, eventually causing the observed bone displacement as the soft tissues decomposed and the gaps between chunks of roofing material filled with sediment. It seems most likely that in either scenario the "burned and salvaged" roof deposit in Structure 5-S would have been created through intentional human agency near the end of the occupation at the site. This supports the inference that "burned and salvaged" roof deposits in Areas 3, 4, 5, and 6 were created during the time of the final emigrations of Pueblo people from the Mesa Verde region and further implies that these structures were in use at that time.

23
In contrast, the tested kivas in the canyon bottom (Areas 1 and 2) contain unburned, salvaged roofs, suggesting that they were abandoned many years before the final emigrations and thus were built even earlier in time. In addition, the bench face of Structure 9-S had been partly dismantled before this subterranean kiva filled with sediment. This suggests that the building stone in this structure was recycled for use in one or more new structures being built nearby, in which case Structure 9-S would have been abandoned during the occupation of the village, rather than at the end of occupation.

24
About 1 to 2 m² of floor area were exposed in each tested kiva at Woods Canyon Pueblo. The total weight of artifacts found on these floors suggests that structures in Areas 3, 4, 5, and 7 were abandoned by people intending to make long-distance moves, such as occurred during the final regional emigrations of the late A.D. 1200s. In contrast, the relative sparseness of the floor assemblages in kivas in Areas 1 and 2 (the canyon bottom) suggests that inhabitants of these structures moved short distances and curated usable artifacts. These data are consistent with the earthen upper lining walls and unburned, salvaged roofs observed in these structures and further support an interpretation of pre–A.D. 1250 occupation. The floor assemblage from Structure 8-S (in Area 6) is also sparse, but this structure has a masonry upper lining wall and its roof had been both burned and salvaged, which suggests occupation in the mid- to late 1200s. Structure abandonment patterns are discussed more fully in "Abandonment and Emigration."

Stratigraphic Evidence

25
Stratified cultural deposits—that is, multiple cultural deposits resting on top of one another—were identified in the canyon bottom. In Nonstructure 1-N, several extramural features rested above and below secondary refuse (trash) deposits. For example, a pit was found above a secondary refuse deposit and a possible extramural stone wall in one excavation unit. In several other cases, extramural stone walls rested on secondary refuse. In Nonstructure 5-N, several distinct layers of refuse were found on top of one another. An analysis of the pottery from the different refuse deposits did not reveal temporal differences. Nevertheless, the presence of stratified trash deposits and the overall pottery signature (see paragraphs 27–33) provide evidence that the canyon bottom was occupied longer than other parts of the site. Long-term and continuous occupation of the canyon bottom is further indicated by the presence of a deeply buried, partly dismantled structure (Structure 9-S) found under Nonstructure 1-N.

26
In ancient times, Pueblo people often discarded their refuse in nearby abandoned buildings. Therefore, the presence of refuse in the fill of some structures indicates that people were still living in the village after these structures were no longer in use. In two kivas, one in the canyon bottom (Structure 1-S) and the other in the upper west side (Structure 7-S), possible refuse deposits were identified above roof fall, but the artifact density in these deposits is much lower than is typical of true middens (see "Artifacts," paragraphs 145–150). If they are cultural deposits, their presence suggests that the two kivas might have been used relatively early and then abandoned, while a portion of the site remained occupied. The low density of artifacts, however, would seem to indicate that refuse was not deliberately discarded in these structures after they were abandoned.

Pottery Dating

27
Although it requires some effort to demonstrate it, we believe that pottery data from Woods Canyon Pueblo indicate that the site was occupied throughout the Pueblo III (A.D. 1140–1280) period. The pottery assemblage from Woods Canyon Pueblo is discussed in greatest detail in "Artifacts." In this section, we provide a summary of the data and arguments developed in "Artifacts" to date the occupation of Woods Canyon Pueblo.

28
Counts and weights of all pottery types identified in the Woods Canyon Pueblo assemblage are presented in Table 2. This table indicates that most sherds in the assemblage were assigned to local Mesa Verde–tradition types characteristic of the Pueblo II and Pueblo III periods. A few sherds were assigned to earlier (Basketmaker III and Pueblo I period) types, including Chapin Gray, Chapin Black-on-white, Moccasin Gray, Mancos Gray, Indeterminate Neckbanded Gray, Early White Painted, and Early White Unpainted. These sherds suggest some form of human activity in the site area between A.D. 600 and 900, but are too rare to indicate occupation during these centuries.

29
Table 3 presents the total count, weight, and percentage by weight of decorated white ware sherds in the Woods Canyon Pueblo assemblage. These data indicate that most such sherds were classified as Pueblo III (A.D. 1140–1280) types, including Mesa Verde Black-on-white, McElmo Black-on-white, and Pueblo III White Painted. Pueblo II (A.D. 900–1140) types, including Mancos Black-on-white, Cortez Black-on-white, and Pueblo II White Painted, are less common, and earlier (Pueblo I/Basketmaker III) types are present only in trace amounts. The remainder of the painted white ware sherds were classified as Late White Painted, a "grouped" type that includes pottery made during either the Pueblo II or Pueblo III periods.

30
We can examine the pottery chronology of Woods Canyon Pueblo in more detail by comparing the proportion of white ware sherds assigned to various formal types (as opposed to informal or grouped types, after Wilson and Blinman [1995*1:35]) in each tested area with proportions of those same types in idealized pottery assemblages developed by Wilson and Blinman (1999*1) using tree-ring-dated sites with short occupation spans. In this model, early Pueblo III assemblages (A.D. 1140–1180) are dominated by McElmo Black-on-white to the near exclusion of Mancos Black-on-white; middle Pueblo III assemblages (A.D. 1180–1225) contain equal amounts of McElmo and Mesa Verde Black-on-white; and late Pueblo III assemblages (A.D. 1225–1280) contain more Mesa Verde Black-on-white than McElmo Black-on-white.

31
At first glance, the data in Table 4 could be taken as evidence that Woods Canyon Pueblo was occupied during the Pueblo II period as well as the Pueblo III period. Specifically, the proportion of white ware sherds classified as Mancos Black-on-white is higher in all areas of the village than would be expected for a pure, Pueblo III occupation. Although one possible interpretation of these data is that much of the site area was occupied during the late Pueblo II period as well, we believe a number of factors rule this out.

32
First, none of the tested kivas at Woods Canyon have characteristics typical of Pueblo II structures. Second, it appears that analysts were biased toward classifying mineral-painted white ware sherds as Mancos Black-on-white. This could have inflated the proportion of sherds assigned to Pueblo II pottery types in the assemblage (see "Artifacts," paragraphs 13–14). Third, recent research using tree-ring-dated pottery collections indicates that Mancos Black-on-white sherds are more common in early Pueblo III (A.D. 1140–1180) assemblages than has previously been believed (see "Artifacts," paragraph 15).

33
When the pottery assemblages from the seven areas at Woods Canyon Pueblo are considered in this light, the moderate percentages of Mancos Black-on-white pottery in most areas (Table 4) do not appear sufficient to support an interpretation of late Pueblo II occupation of the site. Area 6 does contain a high percentage of Mancos Black-on-white pottery, but most of the total weight is from a single carbon-painted jar sherd weighing 94.7 grams. Also, the frequencies of other pottery types in the Woods Canyon Pueblo assemblage suggest that certain areas were occupied for long periods, resulting in assemblages that blend characteristics of multiple pottery periods. Thus, the predominance of McElmo Black-on-white and notable proportions of both Mancos Black-on-white and Mesa Verde Black-on-white in Areas 1–3 suggest that these areas date from the early and middle Pueblo III periods (A.D. 1140–1225); the predominance of Mesa Verde Black-on-white over McElmo and Mancos black-on-white in Areas 4, 5, and 7 suggests that these areas date from the late Pueblo III period (A.D. 1225–1280); and given the small sample and high proportion (by weight) of Mancos Black-on-white deriving from a single sherd in Area 6, it is likely that the occupation of this area also dates from A.D. 1225–1280. Pottery type data thus suggest that the primary occupation of Woods Canyon Pueblo began in the early Pueblo III period (A.D. 1140–1180) in the canyon bottom, and grew to include the canyon rim, east talus slope, and upper west side by the mid–A.D. 1200s.

Summary and Conclusions

34
Woods Canyon Pueblo has a much longer occupational history than Crow Canyon researchers originally suspected. Before our testing program began, surface indications suggested that the pueblo dated from the mid– to late A.D. 1200s. The results of our testing, however, indicate that its primary occupation began sometime in the mid-1100s and lasted until the late 1200s. On the basis of tree-ring dates, architectural styles, structure abandonment mode, stratigraphic evidence, and pottery data, we have assigned each excavated area of the pueblo to one of two temporal components. Areas 1 and 2, in the canyon bottom, have been assigned to an early Pueblo III component, and Areas 3–7, in the other three sections of the site, have been assigned to a late Pueblo III component. We believe that the early Pueblo III component dates from around A.D. 1140 to approximately A.D. 1225, and the late Pueblo III component from A.D. 1225 to the A.D. 1280s. The date range of the early Pueblo III component thus encompasses both the early and middle Pueblo III periods of Wilson and Blinman's chronological scheme, whereas the late Pueblo III component dates to their late Pueblo III period only. The evidence that has been used to assign each area to one of these two temporal components is summarized in Table 5 (which duplicates Table 2 in "Artifacts").

Possible Late Pueblo II Use of the Site (A.D. 1000–1140)

35
No structures, surfaces, or features dating to the late Pueblo II period were found at Woods Canyon Pueblo. However, sherds classified as Mancos Black-on-white, the predominant pottery type in the Mesa Verde region between A.D. 1000 and 1140, were found throughout the site and could be taken as evidence of a late Pueblo II occupation. We have reviewed three reasons why we believe these sherds do not indicate occupation before A.D. 1140. First, our mapping and testing did not reveal any structures that were definitely constructed in an architectural style characteristic of the Pueblo II period. There are a few buildings at the southern edge of the site that could have been constructed in such a style, but because we did not test these buildings, we cannot rule out the possibility that their surface expressions are the result of their having been dismantled, rather than of their having been constructed earlier in the occupation. Second, analytic biases in Crow Canyon's pottery typology procedures may have led to an inflated proportion of sherds being classified as Mancos Black-on-white (see "Artifacts," paragraphs 13–14). Third, Mancos Black-on-white sherds are actually more common in early Pueblo III sites dating between A.D. 1140 and 1180 than has previously been believed (see "Artifacts," paragraph 15). If in fact Mancos Black-on-white sherds were still being deposited in significant numbers between A.D. 1140 and 1180, one would need to observe a much higher proportion of such sherds than are present in the Woods Canyon Pueblo assemblage to infer a late Pueblo II occupation.

36
It is possible, however, that limited cultural activities—for example, farming the canyon bottom or using the seep springs—did take place at Woods Canyon during the late Pueblo II period. Early use of the site might have been comparable to the inferred Pueblo II use of Catherine's Site, a small site located on a terrace at the base of a talus slope in Sand Canyon. Although the primary occupation of this site dates from the Pueblo III period, the presence of a small number of Pueblo II sherds in the site pottery assemblage leads Varien (1999*4) to suggest that this site may have been used on a limited and seasonal basis during the late Pueblo II period. Like Woods Canyon Pueblo, Catherine's Site has agricultural features and is located near a (possible) spring. The presence of water and productive land for farming may have drawn people to both sites.

37
White ware jar sherds from the canyon bottom, the most likely location of late Pueblo II activity at Woods Canyon Pueblo, were classified as Mancos Black-on-white much more often than were white ware bowl sherds from this same part of the site (see "Water Control and Subsistence," Table 2). Because sherds of this type were most common during the late Pueblo II period, some of them may have been deposited before A.D. 1140 as a result of limited activities related to farming and water retrieval. However, white ware jars were often decorated less formally than bowls, and this could have led to a higher percentage of white ware jar sherds being identified as earlier types than is the case for white ware bowl sherds. Thus, the artifact evidence for limited use of Woods Canyon Pueblo during the late Pueblo II period is equivocal.

Early Pueblo III Component (A.D. 1140–1225)

38
The primary occupation of Woods Canyon Pueblo began in the mid–A.D. 1100s. Even though no individual structures, features, or surfaces are tree-ring dated to this period, various other lines of evidence argue strongly that occupation between A.D. 1140 and 1225 was centered in the canyon bottom (Areas 1 and 2). Mancos Black-on-white pottery was most prevalent in this section of the site, and McElmo Black-on-white was the most common decorated white ware type, occurring in greater proportions than Mesa Verde Black-on-white. Two kivas that were partly lined with masonry were tested in the canyon bottom and are believed to have been constructed before A.D. 1200. The stratified cultural deposits found in the canyon bottom also suggest long-term use of this section.

39
Several lines of evidence also support the inference that the canyon bottom fell into disuse a few decades before the end of the occupation of the site. First, bowls with continuous exterior painted designs, an attribute that occurs on roughly 15 percent of vessels in post–A.D. 1250 assemblages from the Sand Canyon locality (see "Artifacts," paragraph 20), are nearly absent from the canyon-bottom assemblage at Woods Canyon Pueblo. Second, the roofs of tested kivas in the canyon bottom were unburned and salvaged, and few usable artifacts were left on their floors. Both patterns suggest that inhabitants of these structures recycled roof beams and took usable artifacts with them to new residences built nearby, if not actually within the site boundaries, when they abandoned these structures. Third, the bench face of Structure 9-S in Area 1 was partly dismantled, consistent with the inference that new residences were built with recycled construction material from early Pueblo III structures.

Late Pueblo III Component (A.D. 1225–1280s)

40
By the middle 1200s, occupation had shifted up the slope from the canyon bottom to focus on the canyon rim, upper west side, and east talus slope (Areas 3–7). All lines of evidence from Areas 4, 5, and 7 point to late Pueblo III occupation of these areas, but the evidence from Areas 3 and 6 is mixed. In Area 3, the location, roof treatment, and floor-assemblage data support a late Pueblo III date, whereas the pottery data support a somewhat earlier date. In Area 6, location, kiva architecture, roof treatment, and the small pottery sample generally support a late Pueblo III date, but the floor-assemblage data support an earlier date of abandonment. Although the preponderance of evidence supports assignment of Areas 3 and 6 to the late Pueblo III component, it should be kept in mind that the evidence is less conclusive for these areas than it is for other parts of the village.

41
The data in Table 5 suggest that areas assigned to the late Pueblo III component may have been constructed at different times. The rim complex (in Area 7) appears to have been constructed during the final decades of Pueblo occupation in the Mesa Verde region. This interpretation is based on the number of structures in this complex that are tree-ring dated to the A.D. 1260s and 1270s (Structures 7-S, 10-S, 11-S, 13-S, 20-S, and 21-S) and on the dominance of Mesa Verde Black-on-white in the associated pottery assemblage. Furthermore, the large number of beams left on the ground surface and in rooms of the rim complex suggests that wood was not salvaged from this area. In contrast, the higher proportion of sherds classified as McElmo Black-on-white in Areas 3, 4, 5, and 6 suggests that occupation of these areas began somewhat earlier.

42
The number of usable artifacts left on the floors of kivas in Areas 3–6 suggests that their inhabitants emigrated out of the Mesa Verde region when they abandoned these structures. However, we cannot determine whether structures in Areas 3–6 were abandoned before or at the same time as those in the rim complex. In the absence of suitable probability samples for accumulations studies, interpretation of kiva roof treatment is the only evidence available that potentially relates to the rate of emigration from the village. If usable timbers in the roofs of late Pueblo III kivas were salvaged before the remaining roofing materials were burned, such timbers could have been used by a remnant population for heating or construction, implying a gradual emigration process. In contrast, if the roofs of these structures were partly burned and dismantled as part of a deliberate abandonment ritual, it is possible that timbers were not salvaged but left scattered about, leaving open the possibility of rapid emigration with a minimal remnant population. Evidence relating to the way in which Woods Canyon Pueblo was abandoned is discussed more fully in "Abandonment and Emigration." Regardless of what actually occurred, it seems likely that occupation of all areas assigned to the late Pueblo III component was at least partly contemporaneous and that Woods Canyon Pueblo was a large village and community center for at least a portion of the late Pueblo III period.

References cited | To borrow, cite, or request permission