The Archaeology of Castle Rock Pueblo

¹
The most important consideration in interpreting a site is the time of occupation. Knowing the dates of sites enables archaeologists to detect, analyze, and understand changes in culture through time. The history of occupation at Castle Rock Pueblo was analyzed using both absolute and relative dating techniques, including pottery, architectural, stratigraphic, and tree-ring dating. Archaeomagnetic dating was attempted at the site, too, with samples collected from hearths in Structure 401 (a tower) and Structure 108 (a partly underground masonry room). The archaeomagnetic samples were not datable, however, because of the large percentage of sand and small percentage of clay in the sediment.

²
The chronology of the Mesa Verde region is as follows: Basketmaker II, 1000 B.C. to A.D. 500; Basketmaker III, A.D. 500 to 750; Pueblo I, A.D. 750 to 900; Pueblo II, A.D. 900 to 1150; Pueblo III, A.D. 1150 to 1300. The data indicate that Castle Rock Pueblo was occupied only once, and that one occupation has been firmly dated to the late Pueblo III period. More specifically, habitation began in approximately A.D. 1256 and ended sometime after 1274, when large-scale migration out of the Mesa Verde region was under way. This interpretation of a short occupation span is supported by the small amount of architectural remodeling, the modest amount of refuse, and the lack of formal burials in the excavated portions of the site.

Dating with Pottery

³
The characteristics of the pottery collected, especially the decorated white ware, clearly indicate that the village was built during the late Pueblo III period see Table 1 in "Artifacts." A few sherds from earlier time periods were found, but they were not numerous enough to indicate habitation. The latest type of white ware to be produced in this region, Mesa Verde Black-on-white, was the most abundant decorated white ware type collected from the site as a whole. Seven sherds of this type were found for every one sherd of McElmo Black-on-white, the next-earliest decorated white ware type. Many sherds could be identified as decorated white ware sherds dating from the Pueblo III period but could not be identifed specifically as the Mesa Verde or McElmo types.

⁴
Pottery was also used to calculate the length of time Castle Rock Pueblo was inhabited. Varien (1999*1:107) used an estimate of the total weight of cooking vessel sherds present at the site, along with the number of residence groups, to arrive at a length of habitation of 24 to 41 years. These calculations, however, were based on an early estimate of 13 households in the village instead of the final estimate of 15 to 16 households. Recalculating the occupation span using 16 households results in an estimate of 17 to 34 years (see Table 56 and Ortman's discussion in "Artifacts"), which is within the range of occupation suggested by the tree-ring dates.

Dating by Architectural Style

⁵
The buildings at Castle Rock Pueblo are characteristic of late Pueblo III construction in the Mesa Verde region. All walls observed were built of coursed masonry, and a large percentage of the masonry consisted of pecked blocks. Masonry cross-section types commonly used at the site were double-stone-wide, compound-with-core, and double-stone-with-core. Several towers were built. Kivas and associated rooms were closely spaced. At least one square kiva (Structure 103) was constructed in this village. Few square kivas are known in the Mesa Verde region, and this style appears to have been a late development in the history of Puebloan occupation of the region (see "Architecture").

Dating by Stratigraphy

⁶
Stratigraphy may be used to date past events relative to one another, and at Castle Rock it provided the evidence needed to establish that nearly all the structures were still being used when habitation of the village ended sometime after A.D. 1274 (see "Tree-Ring Dating," below). The evidence consisted of layers of collapsed roofing material resting directly on structure floors (with no intervening naturally deposited sediment) and the absence of refuse above those same layers. If structures had been abandoned and their roofs left intact while other parts of the village were still inhabited, then we would expect sediment to have accumulated naturally on the floors of the abandoned structures. If, on the other hand, the roofs of the abandoned structures had been intentionally collapsed while people continued to live in the village, it is likely that trash would have been dumped on top of the fallen roofs by residents of nearby structures. Because we found little evidence pointing to either of these scenarios, it appears that, for the most part, the structures in the village were abandoned at approximately the same time.

Tree-Ring Dating

⁷
Tree-ring dating was used to more precisely date the construction of specific structures at Castle Rock Pueblo (Table 1). This method of dating is immensely useful to archaeologists because it provides dates that are both absolute and independent. Absolute dates are dates given in calendar years. Independent dates are derived by methods that are totally independent of archaeological context and systematics (Dean 1982*2:376). Tree-ring dating is unlike stratigraphic dating, which by itself can date events only relative to one another. Tree-ring dating is by far the most accurate method of absolute dating (another, less accurate, method being archaeomagnetism), and it allows Southwestern archaeologists to ask and answer detailed research questions regarding building sequences, settlement patterns, momentary populations, and culture change.

⁸
Although tree-ring dating provides calendar dates, not all tree-ring dates indicate the year in which a building was constructed. Techniques and principles used to guide the interpretation of tree-ring dates have been presented by Dean (1982*2) and Ahlstrom (1985*1) and have been summarized by Lightfoot (1994*1:25–26). The basic principles and assumptions used in interpreting the 409 tree-ring dates from Castle Rock Pueblo were, briefly, the following:

• Construction usually occurred soon after trees were cut.
• The latest cluster of cutting dates for a structure indicates that those trees were cut to construct that building.
• Earlier clusters of cutting dates indicate timbers salvaged from earlier buildings.
• Noncutting dates are the result of damage to the outside of the timber and do not reflect the year of construction.
• If there are no clusters of cutting dates for a building, the latest cutting date is the best estimate of when the building was constructed.

⁹
A tree-ring sample was collected at Castle Rock Pueblo whenever a fragment of wood was found that appeared to have at least 20 rings, although dating generally requires 30 to 40 rings. A building can be tree-ring dated only when wood with an adequate number of rings is collected from that building. Table 1 summarizes interpretations of the 409 tree-ring dates available for the site, using the principles and assumptions just noted. Additional discussion of the sequence of construction of the village is included in "Settlement Organization."

¹⁰
On the basis of the available tree-ring dates, it appears that the first structure to be built at the site was Structure 204. This was a partly underground kiva near the southwest edge of the site, thought to have been constructed in A.D. 1256. No associated surface rooms were found near this kiva. Structure 206, a fully underground kiva near the west end of the butte, was constructed in 1260. Structure 205 (a room), as well as additional, undefined rooms west and north of Structure 206 and on top of a boulder to the east, could not be tree-ring dated but are believed to have been built and used at approximately the same time by the same residence group.

¹¹
In A.D. 1261, Structure 302, a kiva on the north side of the butte, was constructed. The same residence group might have built Structures 301 (a butte-top structure), 305 (a tower), 306 (an alcove room), 308 (a room), and 309 (an alcove room). Probably also in this suite of related structures were additional butte-top structures that were recorded in historic photographs and documents but that no longer exist. Structure 401 (a tower), located at the northwest edge of the site, was constructed in 1264. Possibly built at the same time by the same group were a D-shaped enclosure (Structure 408) and the kiva it contained (Structure 402), plus an adjacent room (Structure 407).

¹²
In A.D. 1265, Structure 101 was either constructed or remodeled. This fully underground structure is near the southeast edge of the site. We could find no evidence of surface rooms nearby that might have been built and used by the same group that built Structure 101. In 1266, Structure 105, an oversized kiva, was constructed just south of the butte. The large size of this centrally located kiva suggests that it might have been used communally by members of more than one residence group (see "Public Architecture" in "Architecture"), possibly in addition to being used as a residence. Surface rooms 111 and 113, as well as additional, undefined structures west of Structure 111, might have been used with this kiva.

¹³
In A.D. 1268, Structure 405, a fully underground kiva near the northeast edge of the site, was constructed. This kiva is one of a small cluster of buildings that was probably built as a unit, including another kiva (Structure 406), a one-story room (Structure 403), and an adjacent two- or three-story structure (Structure 404/409/410). The three surface rooms were almost certainly constructed as a unit, because the shared wall between 403 and 404 abuts the continuous south wall. This indicates that the four outside walls were built as a unit, and then the dividing wall was constructed. The site-enclosing walls that abut the northwest corner of Structure 403 and the southeast corner of Structure 404 were built after the rooms, so if these structures were built as a unit in 1268, then the enclosing walls were constructed after that year, which was well into the occupation of the village.

¹⁴
Structure 304, a kiva just north of the butte, may have been the last structure built or repaired at the site; it yielded the latest tree-ring date—A.D. 1274. This was not a cutting date, however, so the date of construction is uncertain. Structure 307/310, a two-story room on a ledge midway up the butte face above this kiva, and Structure 203, a surface room at the west end of the butte, could have been constructed and used by the same residence group. Our attempt to find additional surface rooms in the area east of Structure 304 was unsuccessful.

Summary

¹⁵
The dating evidence available for Castle Rock Pueblo indicates that construction began in A.D. 1256. Additional structures appear to have been built every year or every few years after that until the village was abandoned, sometime after 1274. Because the ancient Puebloans are believed to have migrated out of the Mesa Verde region by the early 1280s (see Robinson and Cameron 1991*1:24), it is inferred that Castle Rock Pueblo was abandoned by the early to mid-1280s. Tree-ring dates suggest that the construction and abandonment of Castle Rock Pueblo probably coincided with the construction and abandonment of Sand Canyon Pueblo (Lipe 1995*1:156–157). Additional discussion of the end of the occupation at Castle Rock Pueblo can be found in "The Final Days of Castle Rock Pueblo."

References cited | To borrow, cite, or request permission
Copyright © 2000 by Crow Canyon Archaeological Center. All rights reserved.