Back

20

Dating Summary

Mark D. Varien

Contents

Introduction

Stratigraphy and Tree-Ring Dates
G and G Hamlet
Lillian's Site
Roy's Ruin
Kenzie Dawn Hamlet
Shorlene's Site
Troy's Tower
Lester's Site
Lookout House
Catherine's Site
Saddlehorn Hamlet
Mad Dog Tower
Castle Rock Pueblo
Summary: Testing Program Tree-Ring Dates
Archaeomagnetic Dating
Archaeomagnetic Dating Results
Statistical Comparison of the Archaeomagnetic Dates
Pottery Dating
Dating with Pottery Assemblages
Attribute-Based Dating
Dating Pueblo III Components Using the Ratio of McElmo Black-on-White to Mesa Verde Black-on-White Pottery

Synthesis: Dating the Tested Sites

Introduction

In this chapter, stratigraphy and the results of tree-ring, pottery, and archaeomagnetic dating analysis are used to identify and date the various components at each of the tested sites. The contextual basis for these dating arguments is presented in greater detail in the individual site-description chapters (Chapters 2-14). The methods used to derive the assemblage-based pottery dates are presented in Chapter 15.

The term "component," as used in this chapter, refers to occupations that resulted in the construction of features or structures or in the deposition of recognizable strata. It is possible that excavations failed to uncover evidence of some occupations and/or that such evidence was uncovered but not recognized. Occupations represented by small numbers of artifacts (for example, a single projectile point or sherds from a given time period that make up 1 percent or less of the total assemblage) without associated strata, features, or structures are not considered components here.

The Testing Program focused on sampling the Pueblo III components at each of the tested sites (several of the sites had evidence of multiple components). In this chapter, the emphasis is on the Pueblo III components; these are ranked from earliest to latest, and an attempt is made to identify the components that were contemporaneous.

Stratigraphy and Tree-Ring Dates

Structures, features, and culturally deposited strata were used to identify a total of 24 components at the 13 tested sites. The stratigraphic relationships between deposits, features, and structures were used to relatively date each component; tree-ring samples were analyzed to provide absolute dates. Of the 13 tested sites, only Stanton's Site failed to yield tree-ring dates. The specific contexts from which tree-ring samples were collected, and detailed interpretations of dates based on those contexts, are described in the individual site chapters.

The basis for the interpretation of tree-ring dates has been presented in detail by Dean (1978a, 1978b) and Ahlstrom (1985b). Dean (1978a:226) discusses the interpretation of tree-ring dates in terms of the "reference event," which is the death of the tree, and the "target event," which is the archaeological event of interest. "Bridging events" are the behaviors that link the reference event and the target events (Dean 1978a:226). Ahlstrom discusses bridging events and target events at greater length and specifies the principles that underlie the interpretation of tree-ring dates obtained for samples recovered from structures (Ahlstrom 1985b:57-59). Lightfoot (1992:219) provides a useful summary of the points made by Dean and Ahlstrom, and much of the discussion that follows derives from Lightfoot's summary.

Cutting dates are those with the suffix B, r, or v, including those with a plus sign (+) appended to the suffix.(1) The arguments for including "v" dates as cutting dates have been presented most thoroughly by Ahlstrom (1985b:38, 1989:364). Particularly important to interpretation are clusters of cutting dates, and Ahlstrom (1985b:59-60) develops criteria for evaluating these clusters. Although his criteria are used when possible in the following discussion, most of the tested sites did not yield large clusters of cutting dates. Therefore, the following discussion also relies on clusters of noncutting dates, which Dean (1978b:148) has argued can be considered as closely approximating cutting dates.

Analyzing the distribution of dates for a structure or site is critical to interpretation. Ahlstrom (1985b) found that stem-and-leaf plots (Hartwig and Dearing 1979:16-19) are a useful way of displaying the distribution of tree-ring dates, and other researchers have followed his example (e.g., Lightfoot 1992, 1994). Lightfoot (1994:26) describes how stem-and-leaf diagrams are used to display tree-ring data:

Stem-and-leaf plots are essentially horizontal histograms that display the general distribution of dates, while allowing one to see the individual dates that produce the distribution. The "stem" is usually displayed as a vertical column of numbers representing centuries and decades A.D., and the "leaves" are the final digits of each date, which form the horizontal bars of the histogram.

Using stem-and-leaf plots, Ahlstrom (1985b:64-76) discusses the importance of analyzing the distribution of dates rather than individual samples. He begins with the example of a distribution produced by an ideal case. In the ideal case, a structure would be constructed entirely with new timbers, all the timbers would be recovered as tree-ring samples during excavation, and every tree-ring sample would yield a cutting date. Actual dating situations rarely achieve this ideal, but the degree to which the actual date distribution approaches the ideal can be used as a guide for interpreting assemblages of dates.

Actual date distributions approach the ideal when they are characterized by a peak at the late end of the range of dates. This peak at the late end--called the "terminal cluster"--represents an approximation of the ideal case when there is reason to believe that the dates that produced it are from new beams procured for the construction event. The left end of the distribution falls away more or less gradually toward the early end of the time scale. This left tail represents the processes that cause the distribution to deviate from the ideal. Erosion, shaping of beams, use of dead wood in construction, reuse, and stockpiling are all processes that result in deviation from the ideal distribution (Ahlstrom 1985b:67-76).

In addition to interpreting construction events, Hantman (1983:123) has argued that site abandonments can be inferred as having occurred shortly after the latest tree-ring date if an adequate sample of dates is present. Responding to Hantman, Ahlstrom noted that for a sample to be adequate it must (1) include dates for the portions of the site built late in the occupation and (2) have a continuous date distribution, particularly near the end of the occupation (Ahlstrom 1985b:79). Because date distributions for the tested sites do not always meet these requirements, they cannot always be used to infer precise abandonment dates.

Stem-and-leaf plots are used below for Testing Program sites with more than 10 tree-ring dates.

G and G Hamlet

Two stratigraphically distinct occupations were documented for this site. The stratigraphically earlier (Pueblo II) component is associated with post-and-adobe surface rooms and an earth-walled pit structure. Many of the posts in the post-and-adobe rooms were burned, and these provided many of the dates for the site. The later (Pueblo III) occupation is associated with a masonry roomblock and a masonry-lined kiva. Two spatially discrete middens, one associated with each component, are also present.

Forty-three tree-ring dates were obtained for G and G Hamlet, and all are interpreted as being associated with the earlier Pueblo II component (Figure 20.1). The distribution of dates has a terminal cluster in the A.D. 1052-1067 period. Only four of these dates, however, are cutting dates: A.D. 1059, 1064 (two samples), and 1065. The A.D. 1059 date was yielded by a sample collected from the inner-periphery sampling stratum fill. The two A.D. 1064 dates and the 1065 date came from samples in the fill of the Pueblo III kiva (Structure 1); these samples are interpreted as wood that was originally used during the Pueblo II occupation, but was later reused in kiva construction. These cutting dates indicate that construction of the Pueblo II component likely dates to sometime during the A.D. 1059-1065 period.

Three noncutting dates are later than the latest cutting date: A.D. 1066, 1067, and 1083. The A.D. 1083 date is from a post associated with the Pueblo II component. Thus, occupation during the Pueblo II period is interpreted as having lasted until at least A.D. 1083.

Lillian's Site

Lillian's Site also has two stratigraphically distinct components. Like G and G Hamlet, the stratigraphically earlier component consists of post-and-adobe surface architecture, whereas the stratigraphically later component consists of masonry surface architecture (a roomblock and a tower) and a masonry-lined kiva. A well-defined midden is also present. Pottery found at the site suggests that earlier components may be represented, but no architectural features associated with these occupations were uncovered during test excavations.

Sixty-five tree-ring dates are associated with the Lillian's Site occupation (Figure 20.2). Fifty-nine of these are for samples collected from the masonry-lined kiva, Structure 1. Four dates are from a burned room in the post-and-adobe roomblock, Structure 3, which was beneath the masonry roomblock. The final two dates are from samples associated with a post hole on the periphery of the site.

The stem-and-leaf diagram approximates the ideal distribution, with the terminal cluster of dates (18 dates) between A.D. 1200 and 1214. All of these dates are for samples collected from the masonry-lined kiva, Structure 1. Of the 65 dates, only four are cutting dates: A.D. 1207, 1211, and 1212 (two samples). A single noncutting date of A.D. 1214 is the latest date for the site.

Structure 1 appears to have burned at abandonment, but only portions of the burned roof--mostly small beams and closing material--were found during excavation. Larger beams that were still serviceable apparently had been removed before the roof burned. If the burned wood submitted for tree-ring analysis really was material used in roof construction, as inferred, then the latest cutting dates suggest that Structure 1 was built shortly after A.D. 1212, with reused or stockpiled beams included in the roof.

There are no cutting dates for the stratigraphically earlier occupation. The four dates obtained for samples from the fill of an earlier post-and-adobe surface room, Structure 3, are noncutting dates of A.D. 499, 543, 560, and 1075. It is possible that this room dates to the Basketmaker III period, but associated pottery suggests that the A.D. 1075 date may more accurately reflect the period of construction and use. Like the post-and-adobe rooms at G and G Hamlet, the post-and-adobe structures beneath the masonry roomblock at Lillian's Site are interpreted as dating to the late Pueblo II period. The earlier dates for Structure 3 may represent the use of dead wood in this post-and-adobe room; Ahlstrom (1985b:640-642) discusses how dead wood often appears to have been incorporated into jacal walls.

There is support for this dating of the late Pueblo II occupation in the overall distribution of dates. A small cluster (16 dates) of noncutting dates exists between A.D. 1048 and 1095. The fact that the dates cluster suggests that these samples may not be missing many rings. This cluster may represent wood harvested during the second half of the eleventh century. With the exception of the single sample that yielded an A.D. 1075 date, all of these samples are from the fill of the later masonry-lined kiva, Structure 1. This suggests that wood initially used during the Pueblo II occupation was subsequently salvaged and reused during the Pueblo III occupation that dates to the early A.D. 1200s.

Finally, two noncutting dates of A.D. 899 and 964 were obtained from wood found next to a post hole located on the periphery of the site. These are interpreted as dating to the Pueblo II occupation of the site.

Roy's Ruin

The only architecture identified by testing of this site was a masonry roomblock, a masonry tower, and a masonry-lined kiva. A substantial midden associated with these architectural units was also found. A thin stratum of cultural material lies beneath the roomblock. This stratum, together with a small amount of Pueblo II pottery, indicates that there may have been an earlier use of the site, but it appears to have been limited compared with the use during the later occupation.

Twenty-one tree-ring dates were obtained for the site (Figure 20.3), all from samples in the fill of the masonry-lined kiva, Structure 1. Only two are cutting dates, and both of these date to A.D. 1213. A single noncutting date of A.D. 1223 is the only date later than these cutting dates. Like the tree-ring samples from the fill of the kiva at Lillian's Site, the samples from Roy's Ruin are interpreted as smaller beams and closing material that burned after the larger beams had been salvaged. Thus, the kiva is believed to have been constructed in the early A.D. 1200s, shortly after A.D. 1223.

The earlier noncutting dates--including a cluster in the A.D. 1160s--are difficult to interpret. These may be from specimens that would have dated to the A.D. 1200s if all of the rings had been present. The possibility exists, however, that the earlier dates are from wood associated with an earlier use of the site or from wood from a nearby site that dates to an earlier period. This potentially earlier wood may have been salvaged for reuse during the occupation of the site that dates to the A.D. 1200s.

Kenzie Dawn Hamlet

Stratigraphically, Kenzie Dawn Hamlet was the most complex site excavated during the Testing Program. Testing resulted in the identification of nine pit structures, seven surface structures, 25 extramural features, and abundant secondary refuse. These structures and features are believed to date to one of two main periods of occupation identified at the site. The first is a Basketmaker III habitation that dates to the late A.D. 600s. The second is a long, relatively continuous use of the site as a habitation that spans the late Pueblo II through early Pueblo III periods. In addition to these two main occupations, there may have been limited use of the site during the Pueblo I, early Pueblo II, and middle Pueblo II periods, but the evidence for this is equivocal. Finally, there is good evidence for limited use of the site in the A.D. 1200s, after it was abandoned as a habitation.

The earliest occupation dates to the Basketmaker III period. A surface room, Structure 9, and a pit structure, Structure 16, are associated with this occupation. The Pueblo I-Pueblo II limited use is inferred on the basis of pottery dating to these periods; no architecture was identified. The Pueblo II-Pueblo III habitation began with the construction of a pit structure, Structure 8. Post-and-adobe walls and associated surfaces are believed to be part of a roomblock associated with this structure. The Pueblo II post-and-adobe walls are covered by Pueblo III deposits, including a surface room, Structure 5. Structure 8 was truncated by a Pueblo III masonry-lined kiva, Structure 4. Three small, earth-walled pit structures (Structures 12, 13, and 14) are difficult to date. Stratigraphically, they are beneath a masonry roomblock that includes Structures 1, 2, 3, 7, and 15, so they are clearly earlier than this roomblock. Pottery and tree-ring dates from the fill of the small pit structures suggest that they are later than the building episode associated with Structure 8, but earlier than the building episode associated with Structures 4 and 5. The masonry roomblock was extensively remodeled; Structures 1, 2, 7, and 15 appear to have been part of the original construction, but Structure 3 was associated with the remodeling episode. Two masonry-lined kivas, Structures 6 and 11, and a pit structure, Structure 10, are associated with this roomblock, probably when Structures 1, 2, 7, and 15 were in use. Structure 3 may have been used after these pit structures were abandoned, and the remodeling that produced Structure 3 may indicate a change in site use from a habitation to a field house.

Figure 20.4 presents the stem-and-leaf plot of the 42 tree-ring dates for Kenzie Dawn Hamlet (dates for individual structures are presented in Chapter 5). This date distribution departs from the ideal to a greater degree than any of the other tested sites, the result of the complex use history of Kenzie Dawn Hamlet.

The Basketmaker III component can be dated by burned wood from Structure 9. A cutting date of A.D. 659 and two noncutting dates of A.D. 649 and 665 place the construction of Structure 9 in the late A.D. 600s, probably between A.D. 660 and 670. The noncutting dates in the A.D. 500s were yielded by samples from the fill of a Pueblo III pit structure, Structure 13. These dates may be from wood that was originally used during the Basketmaker III occupation, then reused during the construction of Structure 13, or they may be from wood that--had the entire ring sequence been present--would have dated to later in the Basketmaker III occupation.

The Pueblo II-Pueblo III architecture at the site included post-and-adobe walls and associated surfaces. Burned posts from the walls yielded noncutting dates of A.D. 986, 1024, 1036, and 1061, which indicates that the walls were constructed sometime after A.D. 1061. Two cutting dates of A.D. 1081 and 1087 were provided by samples collected from the fill of Structures 14 and 15. This wood is interpreted as material originally used during the Pueblo II occupation, then reused in Structures 14 and 15. Thus, construction during the Pueblo II occupation may date to the A.D. 1080s, approximately the same period as the Pueblo II occupations at G and G Hamlet and Lillian's Site. The Pueblo II occupations of all of these sites included the use of earthen architecture.

Pottery and tree-ring dates in the fill of Structures 12, 13, and 14 suggest that these small pit structures date to the middle A.D. 1100s. Five noncutting dates of A.D. 1110 (two specimens), 1128, 1129, and 1142 are from wood interpreted to be associated with the Pueblo III use of the site. One sample dating to A.D. 1110 was collected from Structure 14, and the sample dating to A.D. 1142 was collected from Structure 13. Structures 12, 13, and 14, which are similar in terms of their form and stratigraphic placement on the site, are interpreted as dating to sometime after A.D. 1142.

The remaining three samples are from contexts associated with the stratigraphically later Pueblo III component. Noncutting dates of A.D. 1110 and 1129 are from samples collected from Structure 10, and an A.D. 1128 sample was associated with Structure 15, a room that overlies Structure 14. Structures 10 and 15, and the other architectural units associated with this building episode (Structures 1, 2, 6, 7, and 11), almost certainly date to the late A.D. 1100s and early A.D. 1200s, an inference made on the basis of their associated pottery.

The final building episode on the site was the roomblock remodeling that produced Structure 3 and the construction of an intrusive feature in Structure 1. These events appear to date to the A.D. 1200s, and they may represent use of the site after the kivas (Structures 6 and 11) were abandoned and had started to fill. This in turn may signal a change in site use from habitation to field house.

Shorlene's Site

Shorlene's Site, like Kenzie Dawn Hamlet, is stratigraphically complex. The latest component consists of a masonry roomblock, a masonry tower (Structure 3), a masonry-lined kiva (Structure 1), and a substantial midden. Structure 1 was constructed inside two earlier pit structures, Structures 4 and 5. Another earlier pit structure, Structure 2, was also found. On the basis of these architectural remains, two components--one Basketmaker III and the other Pueblo III--are confidently identified.

Of the six tree-ring dates for Shorlene's Site, five were obtained from wood used during the Basketmaker III occupation. Three of these are from a burned pit structure, Structure 4, and they consist of two noncutting dates of A.D. 647 and 656 and a cutting date of A.D. 655. Structure 4 is therefore interpreted as having been constructed in approximately A.D. 660.

Noncutting tree-ring dates for Structures 2 and 5 date to A.D. 636 and 653, respectively. These dates may indicate that these pit structures also date to the A.D. 600s and are part of the Basketmaker III component.

The only tree-ring date associated with the Pueblo III component is a noncutting date of A.D. 1145. The Pueblo III component therefore dates to sometime after the middle A.D. 1100s.

Troy's Tower

Small amounts of Pueblo II pottery were found at Troy's Tower, but they total only 1 percent of the entire assemblage so the site is interpreted as a single-component, Pueblo III site. A tunnel connects a masonry tower, Structure 1, to a masonry-lined kiva, Structure 2. Two large pits, Structures 3 and 4, are the other architectural features discovered at the site. Structure 3 is interpreted as having been used as a roasting pit before it was abandoned. A small midden is associated with these architectural features.

There are 29 tree-ring dates for Troy's Tower, all from samples collected from the roasting pit, Structure 3 (Figure 20.5). Three of these are cutting dates: A.D. 1265, 1268, and 1271. A noncutting date of A.D. 1271 was also obtained. These dates indicate that Structure 3 was used sometime after A.D. 1271. The large spread in the dates may be due to the use of dead wood as fuel.

Lester's Site

This talus-slope site is interpreted as a single-component, Pueblo III site. Pueblo II pottery is present, but in small amounts. Architectural facilities associated with this component include two masonry-lined kivas (Structures 1 and 3), two cliff-face alcove rooms (one, Structure 2, was tested), an estimated minimum of three surface rooms, and several retaining walls.

Structure 1 burned, and 90 tree-ring dates were obtained from the samples submitted for analysis (Figure 20.6). The distribution is skewed toward the late end, but the largest cluster is not at the terminal end of the date range. Only six cutting dates were obtained: A.D. 1206, 1207, 1209, 1211, 1238, and 1270. All of these cutting dates are "v" dates except for A.D. 1270, which is an "r" date. The latest date for the site is a noncutting date of A.D. 1271.

Although the noncutting dates are difficult to interpret, the distribution of dates for Structure 1 suggests that the majority of the roof was constructed with reused timbers. The largest cluster of dates, 31 samples (34 percent), lies between A.D. 1190 and 1239, with nine of these falling in the A.D. 1230s. All of the cutting dates listed above, with the exception of the A.D. 1270 date, are included within this span. Structure 1 may have been constructed in A.D. 1270 with timbers salvaged from a roof constructed in the A.D. 1230s or early 1240s. Alternatively, Structure 1 may have been constructed in approximately A.D. 1240, then remodeled in the early 1270s.

Lookout House

Lookout House is another site interpreted as a single-component, Pueblo III site. Testing identified three masonry-lined kivas (Structures 1, 2, and 7), two surface rooms (Structures 3 and 4), a subterranean masonry-lined room (Structure 6), several retaining walls, and two middens.

Only four tree-ring dates were obtained for this site. These consist of three noncutting dates (A.D. 1004, 1104, and 1241) and one cutting date (A.D. 1257). The cutting date was provided by a sample collected from ashy secondary refuse at the base of a midden in an abandoned kiva, Structure 7. Given the context in which it was found, this wood was probably used as fuel. Clearly the midden, and much of the site occupation, postdates A.D. 1257.

Catherine's Site

This site has one main component, which consists of two deep pit structures (Structure 1, a masonry-lined kiva, and Structure 2, whose walls were not observed), a masonry roomblock, several retaining walls, and a substantial midden. Pottery from the site indicates that there may have been an earlier component, but no architectural features associated with this earlier component were identified.

Only five tree-ring dates, all noncutting, were obtained for Catherine's Site: A.D. 938, 992, 1016, 1068, and 1111. The samples that yielded these dates were collected from charcoal in the fill of Structure 2 and from the midden. Because the dates are noncutting dates, are few in number, and do not cluster, they are of little use in dating the occupation of Catherine's Site, other than to narrow the occupation to sometime after A.D. 1111.

Saddlehorn Hamlet

Saddlehorn Hamlet is interpreted as a single-component, Pueblo III site; Pueblo II pottery is virtually absent. Architectural remains include a masonry-lined kiva that burned (Structure 1), masonry rooms adjacent to the kiva, two masonry rooms in an alcove above the kiva, and two structures on the cliff above the rest of the site. A substantial midden is associated with these architectural features.

Sixty-one tree-ring dates were obtained for Saddlehorn Hamlet; 59 samples are from the burned kiva, Structure 1, and two are from the midden (Figure 20.7). The distribution of dates is skewed toward the late end; however, a single noncutting date postdates the latest cutting date by 19 years. Cutting dates for the kiva include 18 samples that date between A.D. 1205 and 1232. There are four cutting dates at A.D. 1210 and seven at A.D. 1231. The latest noncutting date for the kiva is A.D. 1256. Kiva construction may have taken place in either A.D. 1210 or 1231, with subsequent remodeling producing the later dates.

It is also possible that all of the dates before A.D. 1256 represent reused beams. Including noncutting dates, 42 dates for Structure 1 cluster between A.D. 1190 and 1232. Ahlstrom, in his study of tree-ring dating, notes that there is a tendency for beams that were procured together to be used, salvaged, and reused together (Ahlstrom 1985b:72). The roof of Structure 1 at Saddlehorn Hamlet may have been constructed largely from reused timbers that were salvaged from a roof or roofs constructed in the early A.D. 1200s. If the beams were indeed reused, then Structure 1 was constructed sometime after A.D. 1256. The pottery dating supports this interpretation.

The two dates for the midden are a noncutting date of A.D. 1162 and a cutting date of A.D. 1237. The latter is the latest cutting date on the site, and its association with ashy sediments in the midden suggests that the sample was a piece of discarded hearth fuel. The use of this wood as fuel probably postdates the death of the wood in A.D. 1237. Therefore, this cutting date also indicates that the occupation of Saddlehorn Hamlet dates to after A.D. 1237, which in turn supports the argument that the kiva was built after A.D. 1256.

Mad Dog Tower

Mad Dog Tower is also interpreted as a single-component site, despite having a small amount of Pueblo II pottery. A masonry surface room (Structure 1), an earth-walled kiva with masonry pilasters (Structure 2), and a masonry tower (Structure 3) are the architectural features present. The midden at Mad Dog Tower is the shallowest of any of the tested site middens. In terms of site size and architectural features, Mad Dog Tower most closely resembles Troy's Tower.

Only two tree-ring samples from the site yielded dates, both noncutting (A.D. 932 and 1047). Neither is helpful in determining the date of the site.

Castle Rock Pueblo

Castle Rock Pueblo was the largest site investigated during the Testing Program. An estimated 13 to 16 kivas and 45 to 75 surface rooms are present; eight kivas, five masonry rooms, four structures of unknown type, a 9-m-diameter stone enclosure, a tower, a platform on top of the Castle Rock butte, and substantial deposits of secondary refuse were sampled during testing. All the architectural features are associated with a late Pueblo III component. Early pottery is present, but it makes up less than 1 percent of the total assemblage, which indicates that Castle Rock Pueblo is essentially a single-component site.

Tree-ring dates for Castle Rock Pueblo were obtained for a total of 167 specimens (Figure 20.8). Of these, the majority (over 130) came from kivas, 10 were from the tower, four were from the platform on top of the butte, four were from structures of unknown type, 11 were from the midden, and one was from a surface room. Twenty dates are interpreted as cutting dates; these range from A.D. 1165 to 1264. The date distribution shows a terminal cluster in the A.D. 1250s and 1260s. Seventeen (85 percent) of the cutting dates are between A.D. 1253 and 1264, which indicates that the main occupation of the site occurred after A.D. 1253. Earlier cutting dates may indicate reuse of wood from elsewhere or a limited occupation of the site between A.D. 1165 and 1250.

The latest date for the site is a noncutting date of A.D. 1274, yielded by a sample from an abandoned kiva. This is one of the latest dates for the Sand Canyon Project and for the entire Mesa Verde region. It is clear that the occupation of Castle Rock Pueblo continued after A.D. 1274 and that the abandonment of this site was roughly contemporaneous with the abandonment of the locality and the region as a whole.

Summary: Testing Program Tree-Ring Dates

Testing Program excavations produced a total of 535 tree-ring dates, and 63 of these are cutting dates. Sites with the largest assemblages of dates can be interpreted with the most confidence. The best-dated sites anchor the chronology described in the remainder of this chapter.

Basketmaker III components were most confidently identified and dated at Kenzie Dawn Hamlet and Shorlene's Site. Construction at these sites appears to have taken place during the A.D. 660-670 period.

The next occupations that can be confidently dated are the late Pueblo II occupations at G and G Hamlet, Lillian's Site, and Kenzie Dawn Hamlet. At each of these sites, the Pueblo II components were stratigraphically earlier than the Pueblo III components. Post-and-adobe roomblocks were found at each of the sites, and earth-walled pit structures were found at G and G Hamlet and Kenzie Dawn Hamlet. These occupations date between A.D. 1050 and 1100 at all three sites. G and G Hamlet and Kenzie Dawn Hamlet are the best-dated Pueblo II sites; the range of cutting dates at the two sites spans the A.D. 1059-1087 period.

Lillian's Site and Roy's Ruin are the mesa-top unit pueblos that have the most-confidently dated Pueblo III components. Construction at Lillian's Site is interpreted as dating to sometime during the A.D. 1211-1220 period, whereas construction at Roy's Ruin is interpreted as dating to sometime during the A.D. 1213-1230 period.

Occupations after A.D. 1250 are also well dated. The largest assemblage of dates comes from Castle Rock Pueblo. The majority of the construction at Castle Rock is dated to the A.D. 1253-1264 period, with some occupation after A.D. 1274. Construction of a kiva at Lester's Site is interpreted as taking place either in approximately A.D. 1240, with remodeling in the 1270s, or in the A.D. 1270s, with reused timber. A roasting pit at Troy's Tower was used sometime after A.D. 1271. The Saddlehorn Hamlet dates are also subject to alternative interpretations, but the kiva on this site was probably built sometime after A.D. 1256. A single cutting date for Lookout House indicates that this site was occupied sometime after A.D. 1257.

Archaeomagnetic Dating

The results of archaeomagnetic dating for all the sites are summarized in Table 20.1. Although the reports from the archaeomagnetic laboratory often provide a series of possible dates, Table 20.1 lists only those date ranges that are closest to the dates suggested by the results of tree-ring and pottery analyses. In the following section, the archaeomagnetic date ranges are used (1) to improve the dating arguments for sites with poor tree-ring results and (2) to date the site abandonments more accurately.

Archaeomagnetic dating results were also submitted to William Deaver at the Department of Geosciences, University of Arizona, to statistically compare the samples to one another. The goal of this analysis was to relatively date the samples and to determine which are probably contemporaneous (Deaver 1988, 1989). The final section of the dating discussion reports the results of this statistical analysis.

Archaeomagnetic Dating Results

The samples from G and G Hamlet are from Structure 1, the Pueblo III kiva, and from a courtyard hearth, also associated (stratigraphically) with the Pueblo III component of the site. The Pueblo III component was not dated with tree-ring dates. The archaeomagnetic date for the courtyard hearth does not conform to any known occupation of the site, and certainly not to the Pueblo III occupation. The archaeomagnetic sample from Structure 1, however, does indicate that the occupation of this kiva, and the Pueblo III component by association, dates to sometime before A.D. 1250.

Lillian's Site is one of the better tree-ring-dated sites, with construction placed in the A.D. 1211-1220 period. The archaeomagnetic date for the kiva hearth supports this interpretation and indicates that the site was abandoned sometime before A.D. 1250.

Structure 6 at Kenzie Dawn Hamlet is the masonry-lined kiva associated with the last major building episode at the site. Tree-ring dates are not available for this building episode; however, an earlier Pueblo III building episode is tree-ring dated to sometime after A.D. 1142. Thus, the late building episode also postdates A.D. 1142. The archaeomagnetic date for the masonry-lined kiva provides general support for this interpretation, indicating that Structure 6 was used between A.D. 1150 and 1400.

The two dates for Troy's Tower are interesting. Use of Structure 3, a roasting pit, is tree-ring dated to sometime after A.D. 1271. The kiva, Structure 2, was not tree-ring dated, but the results of archaeomagnetic sample analysis indicate that this structure dates to the A.D. 1225-1325 period and is roughly contemporaneous with the roasting pit. The archaeomagnetic date for the roasting pit (A.D. 1175-1400) indicates that this feature was used at the same time as, or perhaps slightly earlier than, the kiva.

Structure 1 at Lester's Site is tree-ring dated to either A.D. 1240 or sometime after A.D. 1271. The archaeomagnetic date for the kiva hearth (A.D. 1275-1650) supports the latter interpretation and indicates that the abandonment of this structure dates to sometime after A.D. 1275. Abandonment of the region is interpreted as having taken place sometime before A.D. 1300. Thus, if Structure 1 was constructed in the early A.D. 1270s, it could have been used for a maximum of about 30 years; if, however, the structure was built around A.D. 1240, it could have been used for a maximum of 60 years.

Lookout House has few tree-ring dates; however, one cutting date obtained from a piece of charcoal in the midden indicates that the occupation dates to after A.D. 1257. The archaeomagnetic date (A.D. 1300-1500) for this site suggests that abandonment may not have taken place until A.D. 1300. This date is interesting not just because it helps us date Structure 1, but because it suggests that the final abandonment of the region may not have taken place until A.D. 1300.

Tree-ring samples from Catherine's Site did not yield any useful dates. The archaeomagnetic date for Structure 2 (A.D. 1200-1325) therefore improves the dating of the Pueblo III component of this site, by indicating that occupation occurred sometime after A.D. 1200.

The archaeomagnetic dates for Saddlehorn Hamlet do not improve the chronological interpretation made on the basis of tree-ring dates, and only support the interpretation that the site was occupied during the A.D. 1200s. Mad Dog Tower tree-ring dates were scarce and of limited relevance in dating the site. Unfortunately, the archaeomagnetic date for this site does not improve the dating, except to indicate that the site could date to the latest portion of the known Puebloan occupation in the Mesa Verde region. Finally, archaeomagnetic samples from Stanton's Site did not produce interpretable results.

Statistical Comparison of the Archaeomagnetic Dates

Deaver (1988, 1989) presents the methods for the statistical analysis of archaeomagnetic dates. Statistical analysis is designed to relatively date the samples and to assess the possible contemporaneity of the samples. The analysis included samples from Green Lizard and Sand Canyon Pueblo, as well as the samples from the tested sites.

Results of Deaver's statistical analysis are difficult to interpret because the variation between samples is continuous, producing overlap between the samples. The analysis placed the samples into four groups (Table 20.2); because of the overlapping variation, some samples appear in more than one group (Deaver 1992). The potentially earliest samples begin with the first sample listed in Group 1; the table entries then move toward the latest sample, which is the last sample listed in Group 4.

Comparing the results of this analysis to what we know about the sites on the basis of tree-ring dates raises some questions about the precision with which the statistical results from the archaeomagnetic data can be interpreted. For example, the relative sequence of the samples from earliest to latest does not appear to be correct when compared with the results of tree-ring analysis. It is also clear that not all of the archaeomagnetic samples that analysis shows to be potentially contemporaneous are in fact contemporaneous. For example, the sample from Lillian's Site and the sample from Structure 501 at Sand Canyon Pueblo are in the same early archaeomagnetic group, yet the kiva at Lillian's Site is tree-ring dated to the A.D. 1207-1220 period, with an abandonment before A.D. 1250, and Structure 501 at Sand Canyon Pueblo is dated to approximately A.D. 1250, with an abandonment coterminous with the abandonment of the site by approximately A.D. 1290.

Groups 1 and 2 are statistically so similar that it is possible that all of the samples in both groups are contemporaneous (Deaver 1992). The most distinct break in the samples is between Groups 2 and 3. This analysis suggests that the samples in Group 3 are likely to be later than the samples in Group 1. The samples in Group 3 might be later than the samples in Group 2, but given the overlap, it is possible that the sample from Kenzie Dawn and the sample from the roasting pit at Troy's Tower are contemporaneous with all of the samples in Group 2.

Statistically, the strongest statement that can be made about this set of archaeomagnetic samples is that the samples from the kivas at Lookout House and Lester's Site and the sample from the courtyard hearth at G and G Hamlet are later than the remainder of the samples. This interpretation may be correct--the kivas at Lookout House and Lester's Site do have late tree-ring dates. Still, these results are surprising in some respects. First, the samples from Sand Canyon Pueblo are thought to be from structures whose abandonment is coterminous with the abandonment of the site, the locality, and the region. Thus, it is surprising that the analysis of archaeomagnetic dates suggests that the abandonment of the sampled structures at Lookout House and Lester's are later than the abandonment of Sand Canyon Pueblo structures. Second, the Pueblo III occupation at G and G Hamlet is thought to date to the late A.D. 1100s to early 1200s. It is possible that the sampled hearth represents a late reuse of the site after it was abandoned as a habitation.

There are several important questions we had hoped to answer with the analysis of archaeomagnetic dates. For example, we wanted to know if the two samples from Troy's Tower were contemporaneous. The statistical analysis indicates that it is indeed likely that they were. The strength of this interpretation could be questioned, however, because the analysis also suggests that other samples were contemporaneous when, on the basis of other dating methods, it is clear that they were not. The discrepancy between archaeomagnetic dating and tree-ring dating calls into question the accuracy of the methods used to relatively date structures using archaeomagnetic dating results.

It may be that the temporal resolution needed to answer the chronological questions raised by the Testing Program are too fine-grained for this method; we are trying to discriminate between sites that may be separated by only a few decades. The samples were collected from sites that almost certainly were occupied within 100 years of each other, and conceivably within 40 years of each other (that is, if the mesa-top unit pueblos were abandoned at A.D. 1240 and the region was abandoned by A.D. 1280--not an impossible scenario). This time frame is within the scope of "archaeological contemporaneity" for many parts of the world. This study may be telling us that, at the temporal scale of 40 to 80 years, these samples are indeed contemporaneous. Discriminating between one or two decades may be too difficult.

Pottery Dating

In Chapter 15, which includes a detailed discussion of pottery dating, assemblages were used to identify components and to estimate the intensity of occupation for each component. Dating with assemblages using the methods outlined by Wilson and Blinman (1991) provides a broad temporal range for each component. In the first part of this section, the assemblage-based dating developed in Chapter 15 is summarized. In the remainder of this section, an attempt is made to further refine the dating of the Pueblo III components at each site. First, an attribute-based approach developed by Hegmon (1991) is summarized. Next, the ratio of McElmo Black-on-white to Mesa Verde Black-on-white pottery is used to rank the Pueblo III components from earliest to latest.

Dating with Pottery Assemblages

In Table 20.3, date ranges are assigned to each of the components on the basis of a comparison of the tested-site pottery with the idealized pottery assemblages identified by Wilson and Blinman (1991). The assignment of site function in this table is made on the basis of the types of architectural features present at the sites.

Assemblage dating is successful in identifying components and placing them in a broad temporal framework. If McElmo Black-on-white is the dominant white ware type, and Mesa Verde Black-on-white is absent, the pottery assemblage is interpreted as dating between A.D. 1140 and 1180 (Wilson and Blinman 1991:47). Wilson used a tree-ring-dated site to argue that the first appearance of Mesa Verde Black-on-white pottery was no earlier than A.D. 1180, and perhaps as late as A.D. 1200 (Wilson 1991:734). Thus, if Mesa Verde Black-on-white is present in the Pueblo III components of the tested sites, it is interpreted as dating to after A.D. 1180. Wilson and Blinman's assemblage dating divides the late Pueblo III into the A.D. 1180-1225 and 1225-1300 periods; one of the differences in the pottery assemblages in these two periods is the presence of Dolores Corrugated in the earlier period and Mesa Verde Corrugated in the later period. Dolores Corrugated is not recognized in the Crow Canyon analysis system, so applying Wilson and Blinman's assemblage dating was only partly successful. This was especially true for the mesa-top unit pueblos that were tested. It was difficult to determine whether these sites fit into Wilson and Blinman's A.D. 1180-1225 or A.D. 1225-1300 periods (Wilson and Blinman 1991:47).

Attribute-Based Dating

Hegmon (1991) examined sites that were tree-ring dated to build a refined pottery chronology for Pueblo III decorated white wares. She chose sites that had evidence of the least amount of reoccupation, and she carefully selected proveniences from multiple-component sites to reduce mixing. Hegmon (1991:19) found four attributes to be the most sensitive indicators of stylistic change in Pueblo III decorated white wares: (1) any framing line, (2) thick-thin framing lines, (3) exterior design, and (4) hatch-solid. Examining pottery from the tree-ring-dated sites, she found that use of these attributes increased over time. The percentages of these four attributes are summed in Figure 20.9 to create indices for the tested sites that Hegmon analyzed, along with Green Lizard and Sand Canyon Pueblo.

Figure 20.9 illustrates that the pottery from Lillian's Site and Roy's Ruin is most similar. These sites are tree-ring dated to the A.D. 1211-1230 period. The pottery from Saddlehorn Hamlet, Castle Rock, Lester's Site, and Green Lizard falls into another group. On the basis of tree-ring dates, Castle Rock Pueblo and Lester's Site are believed to date to sometime between A.D. 1250 and 1280. The tree-ring dates for Saddlehorn Hamlet could be interpreted in a number of ways, with possible construction events in A.D. 1210, 1231, or sometime after A.D. 1256. Hegmon's pottery analysis supports the interpretation that the A.D. 1256 date most accurately dates the construction and use of Saddlehorn Hamlet and that earlier tree-ring dates are from reused beams.

The results of tree-ring dating suggest that Saddlehorn Hamlet, Castle Rock Pueblo, and Lester's Site were partly, if not fully, contemporaneous with Sand Canyon Pueblo. Lester's Site and Castle Rock Pueblo clearly have occupations that postdate A.D. 1270. Despite this contemporaneity, the pottery from Sand Canyon Pueblo stands out as having the highest frequency of the attributes examined. Thus, the stylistic attributes charted in Figure 20.9 may be affected by factors other than time; Sand Canyon Pueblo pottery may be stylistically distinct because of differences in site size, site use, or social factors that relate to how pottery was made and used at the site.

Hegmon's analysis indicates that the decorated white wares from Lillian's Site and Roy's Ruin, which are tree-ring dated to the early A.D. 1200s, differ from those at Saddlehorn Hamlet, Castle Rock Pueblo, and Lester's Site, which date to the A.D. 1250-1280 period. Hegmon found that her attribute-based dating was more accurate than using traditional pottery types to date sites. Similarly, in a comparison of the two dating methods, Ortman (1995) argues that attribute-based dating has certain advantages over dating using traditional types.

Dating Pueblo III Components Using the Ratio of McElmo Black-on-White to Mesa Verde Black-on-White Pottery

Future analyses will extend Hegmon's attribute seriation to all sites. For this report, we use the ratios of McElmo Black-on-white to Mesa Verde Black-on-white pottery to date the Pueblo III components at each of the tested sites (Table 20.4). The problem with using traditional types is that the Crow Canyon analysis places only a small percentage of the decorated white wares into traditional type categories. Still, the ratios of McElmo to Mesa Verde black-on-white pottery are helpful in dating the Pueblo III components at sites that do not have absolute dates.

This table lists the ratios that were calculated using both weights and counts. Some ratios change dramatically depending on whether weights or counts are used (for example, Castle Rock Pueblo), whereas others barely change at all (for example, Kenzie Dawn Hamlet). As noted in Chapter 15, historic land use has affected sherd size. Therefore, it is probably best to use weights when comparing sites. Ratios using counts are presented for comparisons with sites from other projects for which weights are not reported.

The best way to approach Table 20.4 is to first consider sites that have large numbers of tree-ring dates. Lillian's Site and Roy's Ruin have similar ratios of McElmo to Mesa Verde black-on-white pottery. These two sites have similar tree-ring dates, which place construction in the A.D. 1211-1230 period. White wares from these two sites were also stylistically similar according to Hegmon's analysis. Shorlene's Site yielded few tree-ring dates, but its ratio of McElmo to Mesa Verde black-on-white is similar to the ratios at Lillian's and Roy's; Shorlene's Site is therefore interpreted as being roughly contemporaneous with, or perhaps slightly later than, Lillian's and Roy's.

The ratio of McElmo to Mesa Verde at Troy's Tower is also similar to the ratios at Lillian's and Roy's, but the pottery ratio and the tree-ring dates for this site are somewhat at odds with each other. Tree-ring dates obtained for a roasting pit at Troy's Tower place its use in the A.D. 1270s. It may be that most of the Troy's Tower occupation occurred slightly earlier than the use of this roasting pit. In addition, Troy's Tower has the second-smallest pottery assemblage of any of the tested sites, and this may have affected the ratio of McElmo to Mesa Verde when compared with other sites with larger assemblages. Finally, Troy's Tower may have been a special-use site, occupied seasonally or periodically (see discussions of Troy's Tower in Chapters 7, 15, and 22), and this specialized use may account for the site having a lower-than-expected ratio of McElmo to Mesa Verde black-on-white.

G and G Hamlet and Kenzie Dawn Hamlet have the lowest ratios of McElmo to Mesa Verde black-on-white. These sites are interpreted as having Pueblo III components that were earlier than those at Lillian's and Roy's. The most likely dates for the Pueblo III components at G and G Hamlet and Kenzie Dawn are between A.D. 1180, when Mesa Verde Black-on-white was introduced, and about A.D. 1210, when the occupation of Roy's and Lillian's began. Kenzie Dawn Hamlet shows evidence of the most extensive remodeling and rebuilding of any of the tested sites. On the basis of stratigraphy, pottery, and tree-ring dates, Kuckelman (Chapter 5) argues that occupation may have been continuous throughout the A.D. 1100s.

Sites with ratios of McElmo to Mesa Verde black-on-white that are higher than the ratios at Lillian's and Roy's are interpreted as dating to sometime after A.D. 1230. Lester's Site, Mad Dog Tower, and Lookout House have the next-highest ratios of McElmo to Mesa Verde black-on-white. Lester's Site has a burned kiva with tree-ring dates between approximately A.D. 1240 and 1270. Mad Dog Tower has the smallest pottery assemblage of any site, and, like Troy's Tower, it may have been occupied seasonally or periodically and may have had a different use than the other tested sites. However, the pottery, which dates to the mid- to late A.D. 1200s, is consistent with the architecture and archaeomagnetic dates for the site. Lookout House has a tree-ring date of A.D. 1257 for a sample collected from a midden in an abandoned kiva. On the basis of the pattern revealed at these sites, it appears that assemblages with ratios of McElmo to Mesa Verde greater than 1:7 date to after A.D. 1250.

The remaining sites have much greater quantities of Mesa Verde Black-on-white pottery. Except for Castle Rock Pueblo, this is true whether weights or counts are used to calculate the ratios. Castle Rock Pueblo and Saddlehorn Hamlet are tree-ring dated to between A.D. 1250 and 1280. Catherine's Site and Stanton's Site are not tree-ring dated, but their ratios suggest that they, too, postdate A.D. 1250.

At this time, it is not clear why the proportion of McElmo to Mesa Verde differs so much at sites that postdate A.D. 1250. The proportions range from 1:7 at Lester's to 1:34 at Saddlehorn. It may be that the unburned kiva at Lester's was constructed before A.D. 1250 and that this earlier occupation alters the ratio (the same may be true at Lookout House). However, the differences are so great that perhaps some factor other than time affects the relative frequencies of these two pottery types at sites.

In sum, the ratio of McElmo to Mesa Verde black-on-white pottery can be used to place the Pueblo III components at each of the 13 tested sites into one of three general periods: (1) A.D. 1180-1210: sites with ratios of McElmo to Mesa Verde black-on-white lower than those at Lillian's Site and Roy's Ruin; (2) A.D. 1210-1250: sites with ratios of McElmo to Mesa Verde black-on-white similar to those at Lillian's and Roy's; and (3) A.D. 1250-1280: sites with ratios of McElmo to Mesa Verde black-on-white that are greater than those at Lillian's and Roy's. These ranges indicate the period during which the majority of the site occupation is believed to have taken place. Sites and components are assigned to only one period, although the beginning of occupation or the abandonment may overlap slightly with another period. Table 20.4 presents these pottery dates for the Pueblo III components.

Again, the pottery date for Troy's Tower does not entirely agree with the tree-ring dates, which clearly indicate use of the site in the A.D. 1270s.

Synthesis: Dating the Tested Sites

Together, stratigraphy, tree-ring, archaeomagnetic, and pottery dating were used to identify 24 components at the 13 tested sites (Table 20.3). These components represent occupations that resulted in the deposition of strata or the construction of features and structures. Not included in the analysis were small amounts of pottery from periods not associated with identifiable strata, features, or structures. For example, small amounts of Pueblo II pottery are present at every site (see Table 15.5), but this pottery was so scarce at Troy's Tower, Stanton's Site, Lookout House, Lester's Site, Saddlehorn Hamlet, Mad Dog Tower, and Castle Rock Pueblo that we do not recognize Pueblo II components at these sites.

The Basketmaker III components identified include habitations at Shorlene's Site and Kenzie Dawn Hamlet. Tree-ring dates indicate that these components date to approximately A.D. 660 and 670, respectively. These sites are therefore roughly contemporaneous; however, if the occupations of these two sites were short enough, they could have been sequential. Basketmaker III components are also present at Lillian's Site and G and G Hamlet, but the dating and nature (that is, site use) of these occupations are not clear.

Late Pueblo II components were identified at G and G Hamlet, Lillian's Site, Roy's Ruin, Kenzie Dawn Hamlet, Shorlene's Site, and Catherine's Site. Architectural remains at G and G Hamlet and Kenzie Dawn Hamlet include pit structures and surface rooms, and these sites are interpreted as having been used as habitations during this period. Lillian's Site also has surface rooms that date to this period, although no Pueblo II pit structure was found; this site, too, may have functioned as a habitation. Tree-ring dates for samples collected from the Pueblo II components at G and G Hamlet, Kenzie Dawn Hamlet, and Lillian's Site indicate that these components were roughly contemporaneous and that they date approximately to the A.D. 1060-1100 period. The Pueblo II occupations at Roy's Ruin, Shorlene's Site, and Catherine's Site can be dated to between A.D. 1025 and 1100.

Each of the tested sites had a Pueblo III occupation. The dating information for the Pueblo III components is synthesized and interpreted in Table 20.5. The latest tree-ring cutting and noncutting dates for each site are presented in the first two columns. The archaeomagnetic date is presented next. Tree-rings usually date construction events, whereas archaeomagnetic dates may best be interpreted as the date range when abandonment took place. The pottery date, taken from Table 20.4, is presented next. The core date, in the last column, is the period of most intensive occupation for the Pueblo III component at each of the tested sites, an inference made on the basis of all the other dating evidence combined.

The information presented in Table 20.5 can be used to rank the Pueblo III components from earliest to latest and to identify components that were probably contemporaneous. The stratigraphically earliest component at Kenzie Dawn Hamlet represents the earliest Pueblo III component sampled during the Testing Program, although it is not well dated. The stratigraphically late Pueblo III component at Kenzie Dawn and the Pueblo III component at G and G Hamlet are roughly contemporaneous and are interpreted as dating between A.D. 1180 and 1210.

The Pueblo III components at Lillian's Site, Roy's Ruin, and Shorlene's Site appear to be roughly contemporaneous, dating between A.D. 1210 and 1250. Occupation of Troy's Tower and Mad Dog Tower may overlap with these sites; however, the dating of these two sites is less secure, and clearly there is some occupation at Troy's Tower that postdates A.D. 1270.

A number of sites are interpreted as dating primarily to the A.D. 1250-1280 period. Sites that are most securely dated are those with tree-ring dates: Lester's Site, Lookout House, Saddlehorn Hamlet, Castle Rock Pueblo, and the roasting pit at Troy's Tower. Pottery from Stanton's Site and Catherine's Site indicates that these sites also date to the middle to late thirteenth century. All of these sites with late components almost certainly had at least partly contemporaneous occupations, and they are at least partly contemporaneous with the two sites that have been intensively excavated as a part of the Sand Canyon Project: Green Lizard and Sand Canyon Pueblo. The main occupation at Green Lizard dates to the middle A.D. 1200s (Huber and Lipe 1992), and at Sand Canyon Pueblo, the primary occupation dates between A.D. 1245 and 1290 (see Bradley [1992b, 1993] for tree-ring dates).

1. The following suffixes, provided by the Tree-Ring Laboratory, Tucson, Arizona, are used throughout this publication:
B = bark is present.
r = less than a full section is present, but the outermost ring is continuous around the available circumference.
v = a subjective judgment that, although there is no direct evidence of the true outside on the sample, the date is within a very few years of being a cutting date.
vv = there is no way of estimating how far the last ring is from the true outside; many rings may be lost.
+ = one or a few rings may be missing near the outside whose presence or absence cannot be determined because the series does not extend far enough to provide adequate crossdating.
++ = a ring count is necessary beyond a certain point in the series because crossdating ceases.