Appendix A(2), Pottery Dating in The Sand Canyon Archaeological Project: Site Testing

A(2)
Pottery Dating
C. Dean Wilson and Eric Blinman

Appendix A(2) is adapted from an excerpt from an unpublished manuscript, Ceramic Types of the Mesa Verde Region, a handbook prepared for the Colorado Council of Professional Archaeologists Ceramic Workshop (Wilson and Blinman 1991:44-47). Portions of the manuscript, not including this excerpt, have since been published (Wilson and Blinman 1995). Wilson and Blinman continue to work on assemblage-based pottery dating, and more recent work has resulted in slight modifications of the dating periods presented in this appendix.: Mark D. Varien, Editor

The following summary of pottery chronology reflects our current knowledge of patterns of pottery change in the Mesa Verde region. The dating strategy presented here is based on the evaluation of pottery assemblages, with relatively little reliance on the concept of "diagnostic" types. Our approach relies on the examination of the entire pottery assemblage from a site or provenience for dating inferences (Blinman 1988b).

This appears to result in more consistent and accurate date estimates than either pottery attributes (such as line widths or coil heights) or statistical manipulation. Our approach is not significantly different from Colton's use of pottery periods, and we have had the opportunity to examine a large number of tree-ring dated pottery collections and a decade in which to make errors and to correct them. Please realize that pottery dating schemes will always be imperfect, and further work will surely result in modifications and changes.

What follows are descriptions of temporally distinctive pottery assemblages for the Mesa Verde region. These assemblages are idealized in that they characterize broad time periods and, since pottery change is a continuous process, no single characterization will be perfect for both the beginning and ending of a period. Also, the Mesa Verde region is not a homogeneous entity; for example, at one point in time, red ware frequencies in collections can range from 1 to 25 percent across the region. Another weakness is that there are too few recently described and well-dated pottery collections to construct a high-precision pottery chronology for the entire prehistoric occupation of the San Juan region, and many gaps have to be filled with assumptions and guesses. These gaps are decreasing in number as more excavations are completed and reported, and the weaknesses and errors in the dating scheme are continually being exposed and corrected.

The pottery dating process we advocate consists of comparing the content of a collection with the idealized content of these assemblages. The best fit on all types is the logical best date, but points of difference are the keys to determining whether or not the best date is likely to be accurate. An excess of Plain Gray sherds in a demonstrably Pueblo II survey collection is anomalous and, even in the absence of any "diagnostic" Chapin Gray sherds, the possibility of an underlying Basketmaker III or early Pueblo I component should be investigated. The combination of early and late Pueblo I occupations can result in a collection whose contents resemble that of a middle Pueblo I occupation, but there will be enough discrepancies in content to raise a red flag if the collection is carefully compared with the expected patterns. This doesn't mean that all discrepancies and variability should be interpreted in terms of time periods (site function can influence typological content of collections), but temporal explanations are a good starting place.

One limitation of this technique is sample size. Small collection sizes limit dating inferences by interjecting a significant amount of sampling error. This doesn't mean that date estimates cannot be made, just that the precision possible with small collections will be relatively poor. The crucial question is, Is the absence of a type from this collection due to its absence at the site, or is it simply rare enough that the collection missed it by chance alone? No pre-set sample size will be adequate to cope with all occasions, and the best results occur when the field observer senses that something appears to be amiss and that samples need to be augmented or that collections need to be made from different portions of a site to sort out possible multiple components. Since the distinctiveness of time periods differs, a collection size of at least 400 sherds may be necessary to distinguish an A.D. 860-880 from an A.D. 880-910 occupation, whereas 100 sherds would usually be sufficient to distinguish an A.D. 775-825 from an A.D. 825-860 occupation. Whenever multiple components are suspected, sample sizes must be increased if their detection is to be made with any confidence.

The currently distinguishable dating periods defined for the Mesa Verde region, particularly the Montezuma and Dolores valleys, include the following:

A.D. 200-575: Although pottery may occur at sites on the northern Colorado Plateau as early as A.D. 200 (Dittert et al. 1963; Eddy 1961), pottery assemblages dating to before A.D. 575 are extremely rare. Because of their rarity and the general lack of independent dates, the dating of these early assemblages is very speculative. The earliest pottery assemblages appear to be characterized by the presence of brown wares only. Later (by A.D. 400), red wares and smudged pottery may occur along with brown wares (Breternitz 1986). Polished utility ware types, such as Twin-Trees Gray⁽¹⁾ and Sambrito Utility, are present during this entire phase.
A.D. 575-725: Gray ware types are limited to Chapin Gray, Twin-Trees Gray, Sambrito Utility, and Plain Gray. Chapin Black-on-white is the only white ware type, and few white ware sherds are polished. If red ware sherds are present, they consist solely of Dolores Red.
A.D. 725-775: Chapin Gray and Plain Gray remain the principal gray ware types, although traces of Moccasin Gray have been noted in an extreme minority of collections. Both Chapin Black-on-white and Piedra Black-on-white may be present, and a greater proportion of white ware sherds are polished. Red ware types can include Abajo Red-on-orange, Dolores Red, Abajo Polychrome, and Bluff Black-on-red, although the latter types will be in the extreme minority.
A.D. 775-825: Moccasin Gray sherds are usually, but not always, present as a minority of the gray ware sherds, with Chapin Gray and Plain Gray remaining dominant. Piedra Black-on-white is more abundant than Chapin Black-on-white, and polishing is present on most white ware sherds. Red ware sherds are abundant, exceeding the frequency of white ware sherds in collections from the central and western portions of the region. Abajo Red-on-orange is the most abundant type at the beginning of the period, but black-painted sherds (Bluff Black-on-red) are dominant by the end of the period.
A.D. 825-860: Moccasin Gray becomes two to three times more abundant than Chapin Gray during this period. Mancos Gray may be present but only in trace quantities (aberrant sherds from Moccasin Gray vessels). Piedra Black-on-white and Chapin Black-on-white may be present, and polished white ware sherds are abundant. Bluff Black-on-red is the dominant red ware type, and red wares are less abundant as part of the total assemblage than in the previous period.
A.D. 860-880: Moccasin Gray is the dominant gray ware type, with consistent but minority amounts of both Chapin Gray and Mancos Gray. Piedra Black-on-white is dominant, Chapin Black-on-white is rare, and Cortez Black-on-white may be present as design elements but not in its classic (crackled slip) manifestation. Bluff Black-on-red continues as the dominant red ware type.
A.D. 880-910 (930): Moccasin Gray remains the dominant neckbanded type, but Mancos Gray is a close second. Chapin Gray is consistently present but in low frequencies. Piedra Black-on-white is the most common white ware type, and some sherds of Cortez Black-on-white will be present, but still not in classic form. White Mesa Black-on-white is present in collections from the western region and may be present as traces in the central region. Bluff Black-on-red continues as the dominant red ware type. Sherd temper appears in a minority of both white and red ware sherds. A significant proportion of the red wares contain crushed-sherd temper and are assigned to the McPhee variety of Bluff Black-on-red. Some red ware sherds with strong slips and fine line work will be classified as Deadmans Black-on-red.
A.D. 910 (930)-980: Gray ware sherds are dominated by Mancos Gray and Moccasin Gray, with some Chapin Gray and the introduction of corrugated gray wares (Mancos Corrugated). Corrugated gray wares are in the minority at the beginning of the period and are in the majority by the end of the period. Cortez Black-on-white is the dominant white ware type in the central and eastern region, and White Mesa Black-on-white is the dominant white ware type in the west. Deadmans Black-on-red can be the principal red ware type, but Bluff Black-on-red is still abundant. Sherd temper ceases to form a significant proportion of the red wares but increases its frequency in white ware sherds. Slipping becomes more common in white wares.
A.D. 980-1025: Gray wares are dominated by corrugated sherds, all or almost all from Mancos Corrugated vessels. Mancos Gray and Plain Gray are present, the latter in moderate quantities, but Moccasin Gray is rare to absent. Cortez Black-on white and Mancos Black-on-white are both present, with Cortez Black-on-white dominant in the early portion of the period and Mancos Black-on-white dominant in the later portion. All red ware sherds are from San Juan Red Ware types, and Deadmans Black-on-red vessels account for most or all of the sherds.
A.D. 1025-1100: Gray wares are dominated by corrugated sherds, most of which are from Mancos Corrugated and some of which are from Dolores Corrugated vessels. Plain Gray and Mummy Lake Gray sherds may be present in small quantities. Cortez Black-on-white is usually present, but Mancos Black-on-white sherds are the most common white ware type. Almost all of the Mancos Black-on-white sherds are decorated with mineral paint. Most red ware sherds are from the Mesa Verde region, but traces of Tsegi Orange Ware may be present. Toward the end of this period (A.D. 1075-1100), Dolores Corrugated sherds, organic-painted Mancos Black-on-white sherds, and Tsegi Orange Ware become more common.
A.D. 1100-1140: Dolores Corrugated is the dominant corrugated vessel form, although sherds of both Mancos Corrugated and Mesa Verde Corrugated may be present. Mummy Lake Gray and Plain Gray sherds may be present in small quantities. Mancos Black-on-white is the predominant white ware type, and some examples may be executed in organic paint. McElmo Black-on-white is present in most collections. Red ware sherds are dominated by Tsegi Orange Ware, but not to the exclusion of San Juan Red Ware; White Mountain Red Ware sherds may be present but are rare.
A.D. 1140-1180: Dolores Corrugated is the dominant corrugated type, with some Mesa Verde Corrugated and only traces of Mancos Corrugated. McElmo Black-on-white is the dominant white ware type to the near exclusion of Mancos Black-on-white. Red wares are rare, are dominated by Tsegi Orange Ware polychromes, and may include small amounts of White Mountain Red Ware sherds.
A.D. 1180-1225: Both Dolores Corrugated and Mesa Verde Corrugated are abundant, and white ware sherds include an equal mixture of McElmo Black-on-white and Mesa Verde Black-on-white sherds. Red ware sherds are scarce to absent, and when present, are limited to White Mountain Red Ware.
A.D. 1225-1300: Mesa Verde Corrugated sherds are more abundant than Dolores Corrugated. Mesa Verde Black-on-white is more abundant than McElmo Black-on-white, and the former type increases in abundance toward the end of the period. Red wares are usually absent, but if present, they consist of White Mountain Red Ware.

1. The following types mentioned in this appendix are not used in the Crow Canyon pottery analysis system: Twin-Trees Gray, Sambrito Utility, Dolores Red, Abajo Polychrome, and Dolores Corrugated.