CHAPTER 6: THE PRODUCTION AND USE OF PUEBLOAN UTILITY WARES
As discussed in Chapter 4, understanding the production and use of ancient plain and corrugated pottery provides crucial information on how the pottery interacted with different aspects of its environment during the development and spread of corrugation technology. By production, I refer to the conditions under which the manufacture of utility wares took place. These conditions include how potters learned their craft, technologies involved in the manufacture of utility wares, and aspects of the organization of utility ware production. In documenting the use of utility wares, I am concerned with the kinds and intensity of uses in which the manufactured vessels were employed.
Data on utility ware production and use come from two main sources: (1) analyses of the six assemblages from the Mesa Verde region, and (2) a review and synthesis of published information from the Mesa Verde region and elsewhere in the Puebloan Southwest. Consequently, this treatment focuses mainly on the Mesa Verde region. The extent to which the inferences regarding pottery production and use in the Mesa Verde region data are relevant to other areas of the American Southwest must be determined by future research.
Utility Ware Production
During the production process, utility ware vessels-in-the-making interact with a potentially complex environment including the learning and skill of the producers, other artifacts and technologies employed in utility ware production and competing for the same production resources, and the ways in which utility ware production was organized ( Rogers 1995). How potters learned their craft and the skills needed to master a particular set of techniques can be important aspects of the environment in which utility ware technologies evolved. In the previous chapter, we saw that large-scale patterns of migration and other mechanisms of interaction affected the spread of various corrugation techniques. This is probably also true for patterns of interaction at much smaller spatial scales, such as within and between settlements ( Zedeno 1995). The learning curves required by new pottery-making techniques and the ways in which they mesh with existing techniques and skills can also affect the development and adoption of new technologies. For example, corrugation techniques do not mesh well with production technologies in which vessels are formed with a paddle and anvil, a technique used in the Hohokam region, but not yet documented for the Anasazi. In addition, because pottery production involves time and energy costs, utility ware production must compete with other activities, particularly the production of other kinds of pottery, for potentially limited resources. Finally, the organization of utility ware production, particularly the extent to which utility vessels were produced for exchange and the degree of production specialization by individuals or groups, can affect the development and spread of new technologies. If part or full time specialists produced utility ware for exchange rather than individual households producing these vessels for their own consumption, issues such as efficiency and economies of scale can come into play.
The Transmission of Pottery-Making Knowledge
Ethnographic observations of the transmission of knowledge of how to make pottery among historic and modern Pueblo potters indicates that most transmission occurred vertically from members of one biological generation to the next although not necessarily between biological parents and their offspring ( Bunzel 1929; Guthe 1925; Stanislawski 1977, 1978). Bunzel and Stanislawski note that direct training in pottery making took place primarily within families and between biological generations (members of older generations teaching those of younger generations), and usually involved only the technological aspects of manufacture. Painted designs and other decorative treatments were learned through observation and imitation, which could involve a wide variety of models including whole or fragmentary pieces of the pottery itself. Where pottery was manufactured for exchange (in the historic and modern cases, for sale to non-Pueblo people), potters often copied the work of those who were the most successful. This kind of transmission played an important role in the rapid revitalization of white ware production at Hopi in the 20th century ( Bunzel 1929; Danson 1965).
The archaeological record also provides some clues to the ways in which pottery-making knowledge may have been transmitted during the period of change from plain to corrugated pottery. Currently available evidence comes in two forms-settlement and community patterns, and the patterns of movement between areas with distinctive pottery-making traditions. Different ways of spatially organizing settlements and communities can affect the frequency of interactions among people involved in pottery manufacture, and thus the potential for inhibiting or promoting the transmission of pottery-making knowledge. For example, small, dispersed, more or less self-sufficient settlements that are only weakly associated with other settlements as parts of larger communities offer little opportunity for substantial transmission outside individual families. On the other hand, large settlements composed of many households and families, or tightly clustered smaller settlements participating heavily in a larger community offer many more opportunities for transmission among unrelated or distantly related individuals. In addition, if the rate of long distance movement of people involved in pottery production fluctuated, this could also affect the ways pottery-making knowledge was transmitted. Periods of significantly increased short and long-distance mobility could increase the likelihood that new traits would be introduced into the populations or regions receiving these immigrants.
When neck-banding first appeared in the Mesa Verde region and elsewhere in the northern Southwest during the latter half of the eighth century AD, populations were relatively small and lived in dispersed, single household settlements ( Kane 1986). Under these conditions, transmission of pottery-making knowledge was probably limited mainly to within families, and the spread of new traits occurring primarily through migration and exogamous marriage patterns. The potential for greater transmission increased as settlements grew to include multiple households in the ninth century and particularly when large villages (which may have included people who were not closely related) formed in the latter half of the ninth century in the Mesa Verde region ( Kane 1986; Wilshusen 1991). With the apparent large-scale migrations and coalescence of populations in the southern San Juan basin during the late ninth and early tenth centuries, transmission among previously separate and isolated populations probably occurred. The possible formation of a large, integrated polity centered on Chaco Canyon, which apparently facilitated household mobility ( Lekson 1996a), during the tenth and eleventh centuries, probably also increased the potential for more, extra-familial transmission of pottery-making knowledge.
Pottery-Making Technologies
As discussed in Chapter 5, utility wares, in the Mesa Verde region at least, were all constructed by hand using coils or slabs. No pottery wheel technology existed in the Southwest prior to contact with Europeans ( Peckham 1990). The use of a base mold or puki by Pueblo potters allowed vessels to be rotated during the application of coils, but with insufficient speed to produce thrown pottery. In addition, there is no clear or convincing evidence that a paddle and anvil technology was used to form the early plain utility wares. Simple rolling, pinching, and scraping appear to have been the core techniques used to form utility wares throughout the history of Pueblo pottery manufacture. These same methods were employed to form the painted pottery as well, with the addition of slipping, polishing, and painting techniques. No new or special technologies were required to manufacture corrugated pottery. However, full-body corrugated vessels may have required more skill on the part of the potters than plain or simple neck-banded pottery. This is because the corrugation technique allows less flexibility in the shaping the vessel than when making pottery that is scraped plain. When making a corrugated vessel, one must shape the vessel at the same time the coils are applied. In contrast, plain vessels can be molded and shaped by pressure and scraping at any time before the clay hardens.
The production of utility wares differs from painted wares in terms of the selection and processing of raw materials. It is usually possible to distinguish utility wares from white, red, and polychrome wares on the basis of the texture of the paste alone. Most utility wares posses a very coarse-textured paste in comparison to the painted wares. The relatively coarse texture of utility wares is produced by using large, angular material for temper, and processing the paste in such a way that large pores are relatively common. Both coarse temper and abundant pores improve the ability of a pottery vessel to withstand thermal shock and other forms of catastrophic fracture ( Bronitsky and Hamer 1986; Feathers 1989, 1990; West 1992). Although these textural differences appear very early in the Pueblo pottery-making tradition, in the Mesa Verde region the differences become more pronounced through time as the paste of painted wares becomes finer and the paste of utility wares becomes more consistently and completely coarse-textured ( Breternitz et al. 1974).
Once the vessels were formed, they had to be fired to produce a useful ceramic. Very little published data exists on the firing of utility wares in the Mesa Verde region or elsewhere in the Southwest (e.g., Sullivan 1988). Although large, stone-lined kilns have been found in the Mesa Verde region, restorable utility ware vessels are conspicuously absent from most of these features (Joel Brisbin, personal communication; Fuller 1984; Helm 1973; Winter et al. 1977:14-15). Consequently, we still know very little about how utility ware vessels were fired, or if there were any changes in firing during the development of corrugation. By analyzing features of the pottery itself, we can gain some insight into this aspect of production. Documenting the conditions under which ancient pottery was fired is fraught with difficulty because of the challenges involved in isolating and controlling the differential effects of temperature, atmosphere, and time on the physical responses of pottery ( Feathers 1995). Rather than invest considerable resources in conducting complex and costly analyses that may not produce useful results, I decided to record two simple attributes on all sherds analyzed from the six Mesa Verde region collections that might indicate whether a change in firing regime occurred. If these studies indicate a change in firing took place, then the nature of this change can be more thoroughly investigated in the future. The attributes I recorded include the presence or absence of fire clouds on the interior and exterior surfaces of the vessel, and the patterns in the oxidation or reduction of the surface in relation to the core of the vessel wall.
The surface colors of the gray and white ware pottery produced in the American Southwest were obtained by firing in an atmosphere of limited oxidation. Because this atmosphere is difficult to control in primitive pit and surface firing, zones of over oxidation and over reduction frequently appear on the surfaces and in the cores of ancient pottery and in pottery replicated using similar firing technology. Dark, reduced cores can also occur if firing takes place very rapidly, even if the atmosphere of the kiln is oxidizing (Jim Feathers, personal communication). Thus, systematic changes in the occurrence of surface fire clouds and reduced or oxidized cores can indicate patterned alterations in the firing regimes employed. Although only the presence or absence of surface fire clouds was recorded, I recorded 14 different core color patterns, which are shown in Figure 16.
Figure 16. Diagrams of sherd core color pattern classification used in this study. Dark areas are reduced and light areas are oxidized portions of the sherd cross section. The upper side of each diagram refers to the exterior surface of the sherd, and the lower side is the interior surface.
Table 8 shows the data on the patterns of vessel core oxidation and reduction visible on a freshly broken edge. Table 9 shows the abundance of surface fire clouds in the six gray ware assemblages. Although presence of fire clouds remains fairly constant through time, variation in core reduction exists among the assemblages indicating a possible increase in the prominence or degree of reducing firing atmospheres through time. The abundance of reduced cores increased in the tenth and early eleventh centuries AD, but then appears to decrease again in the middle of the eleventh century. Unfortunately, this pattern is not consistent enough to rule out local variation rather than a clear temporal trend. The only comparable data come from the Dolores River Valley where gray ware jars show an almost ten fold increase in the amount of vitrification during the late ninth century AD ( Blinman 1988b). However, we lack comparable data from the tenth and eleventh centuries.
| Assemblages | Core Color Patterns | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| No Reduced Core1 | Central Core Reduced2 | Completely Reduced3 | Exterior Surface Reduced4 | Interior Surface Reduced5 | Both Surfaces Reduced6 | Totals | ||||||||
| g | % | g | % | g | % | g | % | g | % | g | % | g | ||
| 5MT2193 | 1185 | 51.4 | 686 | 29.8 | 50 | 2.2 | 181 | 7.8 | 191 | 8.3 | 11 | 0.5 | 2305 | |
| 5MT4671 | 3542 | 58.6 | 1339 | 22.2 | 224 | 3.7 | 604 | 10.0 | 253 | 4.2 | 81 | 1.3 | 6044 | |
| 5MT3868 | 4302 | 58.5 | 1260 | 17.1 | 473 | 6.4 | 777 | 10.6 | 490 | 6.7 | 55 | 0.8 | 7358 | |
| 5MT8371 | 569 | 36.4 | 273 | 17.5 | 137 | 8.8 | 365 | 23.4 | 169 | 10.8 | 49 | 3.2 | 1562 | |
| 5MT1786 | 337 | 10.2 | 968 | 29.2 | 1115 | 33.7 | 385 | 11.6 | 499 | 15.1 | 8 | 0.2 | 3312 | |
| 5MT8827 | 1632 | 46.7 | 1027 | 29.4 | 332 | 9.5 | 241 | 6.9 | 187 | 5.4 | 74 | 2.1 | 3493 | |
|
Rounding error has produced small discrepancies in the totals. 1Includes pattern B from Figure 16. 2Includes patterns C, F, L, and M from Figure 16. 3Includes pattern A from Figure 16. 4Includes patterns D, H, and K from Figure 16. 5Includes patterns E, G, and J from Figure 16. 6Includes pattern I from Figure 16. |
||||||||||||||
| Assemblages | Presence of Fire Clouds | |||
|---|---|---|---|---|
| Interior Surface | Exterior Surface | |||
| g | % | g | % | |
| 5MT2193 | 275 | 11.9 | 248 | 10.8 |
| 5MT4671 | 538 | 9.0 | 860 | 14.2 |
| 5MT3868 | 998 | 13.7 | 1662 | 22.7 |
| 5MT8371 | 200 | 12.8 | 526 | 33.5 |
| 5MT1786 | 210 | 6.3 | 449 | 13.5 |
| 5MT8827 | 516 | 14.8 | 626 | 17.9 |
| Totals | 2738 | 11.4 | 4371 | 18.2 |
| Percentages calculated separately for each surface. | ||||
Taken together, these data indicate a possible increase in control over firing regimes through time resulting in higher firing temperatures and more reducing atmospheres, both of which yield harder pottery. If this reconstruction is correct, the adoption of corrugation may have coincided with an increase in vessel hardness which may have made the vessels more brittle, and thus, susceptible to thermal shock and fatigue. However, the data are not conclusive and clearly more research on firing regimes is warranted.
Organization of Utility Ware Production
Although there are a variety of ways of analyzing the organization of pottery production ( Costin 1991; Mills and Crown 1995b; Pool 1992), I am most interested in identifying any changes in the concentration and scale of utility ware production. Did each household produce its own utility vessels, or was there some degree of specialization in production for exchange? Differences in these modes of production organization have implications for how and why changes in gray ware manufacture occurred. Investigations of the organization of pottery production usually involve examination of both direct and indirect evidence. Direct evidence consists of the raw materials, tools, facilities, and by-products of manufacture while indirect evidence includes compositional and formal aspects of the pottery itself.
Differential distributions of direct evidence of utility ware manufacture prior to the ninth century AD in the Mesa Verde region indicates that there may have been some household level specialization in production ( Blinman 1988b; Blinman and Wilson 1992). However, the size of the household sample is exceedingly small and thus potentially unreliable, but indirect evidence, in the form of amounts of nonlocal pottery in household assemblages, also indicates some production specialization ( Blinman and Wilson 1992). After AD 800, direct evidence indicates that gray ware production occurred at most if not all households, and the movement of gray ware vessels within the region appears to have been minimal ( Blinman and Wilson 1992; Glowacki 1995; Glowacki et al. 1995; Pierce et al. 1999; Wilson 1988b, 1991). Thus it appears that over the period of change from plain to corrugated vessels in the Mesa Verde region, utility ware production remained unspecialized and mainly, or entirely, occurred in individual households for their own consumption.
However, outside the Mesa Verde region, evidence suggests that other forms of production organization may have existed. Indirect evidence from Chaco Canyon indicates that import of corrugated vessels manufactured in the Chuska Valley to the west of Chaco increased through time. By the early eleventh century, Chuskan vessels constitute half of all utility ware pottery recovered in Chaco Canyon ( Toll 1991). The substantial increase in the movement of utility ware pottery into Chaco Canyon suggests that certainly the scale and possibly the degree of specialization of gray ware production in the Chuska Valley also increased through time. Unfortunately, studies of direct evidence of pottery production in the Chuska region have not yet been conducted.
Another aspect of the organization of pottery production is the allocation of time and energy resources to the manufacture of different kinds of pottery. Figure 17 shows the average relative frequencies of different pottery wares recovered from sites in the Mesa Verde region by fifty-year intervals from the seventh to the thirteenth centuries. For about half of this time period, gray wares (pottery lacking slip, polish, or paint) account for over 80% of all pottery. From the seventh through the ninth centuries, slipped and painted pottery increased gradually relative to gray wares. Most of this increase resulted from a substantial rise in the abundance of red ware pottery imported from southwestern Utah ( Blinman 1988b; Hegmon et al. 1995, 1997; Lucius and Breternitz 1981; Wilson and Blinman 1995). During the tenth century, the importing of red ware pottery from Utah declined while the relative frequency of locally produced white wares increased dramatically. This change in the relative frequencies of pottery wares also coinsides with the adoption of full-body corrugation in the Mesa Verde region. From the eleventh century on, gray wares, the vast majority of which are from full-body corrugated vessels, and white wares account for roughly 60% and 40% of pottery, respectively, assemblages from the Mesa Verde region. A similar change in the relative abundance of utility and painted wares occurs in other areas of the Southwest, but this pattern is by no means universal. Although I have not systematically compiled data from other areas as I have for the Mesa Verde region, my impression from reviewing the literature is that there is considerable variation in the relative frequencies of utility and painted wares at different times across the Southwest. However there may, in some areas, be a correlation between the adoption of full-body, indented corrugation and an increase in the relative abundance of painted pottery (My thanks to C. Dean Wilson for bring this possible relationship to my attention.).
Figure 17. Change in the relative frequency of pottery wares through time in the Mesa Verde region compiled from published sources.
Because these values are relative frequencies, a change in the absolute frequency of one ware category necessitates a change in the other categories. Consequently, it is not clear from the ware data alone whether we are documenting an increase in white wares, a decrease in gray wares, or some combination of the two. In addition, these data record the amount of pottery that accumulated at sites through discard, which is not necessarily the same as the amount of pottery produced. Changes in the use-life and discard rate of pottery in one ware category would also affect the relative frequencies of wares recovered from archaeological deposits. Changes in the size of vessels in different ware categories can also affect the relative abundance of pottery wares in archaeological assemblages. However, larger vessels require more production investment per pot than smaller vessels, so the potential effects of changes in vessel size do not compromise our ability to track the allocation of production resources among different kinds of pottery.
Data on vessel forms and pottery accumulation rates indicate that, in the Mesa Verde region at least, the changes in the relative frequencies of pottery wares resulted from both increased white ware production and decreased gray ware accumulation. Vessel form data can further inform on production investments because certain forms were usually made in one ware category or another. Both the variety and relative abundance of different vessel forms change between the seventh and thirteenth centuries ( Blinman 1988a, 1988b; Pierce and Varien 1999; Wilson 1988b, 1991). Most of the different vessel forms produced in the Mesa Verde region had been developed by the late ninth and early tenth centuries. After the tenth century, few new vessel forms were added, but the relative frequency of different forms continued to change. In addition, the ware category in which different vessel forms were made also changed through time. The most important change in the proportions of different vessel forms for this discussion is a gradual increase in the abundance of bowls (Figure 18). The vast majority of these bowls were slipped and painted. Consequently, an increase in bowl production involved an increase in white ware production as well. Other vessel forms, such as ollas and seed jars (see Figure 19 for diagrams of different vessel forms), which had been produced mainly as plain gray wares prior to eleventh century, were mostly slipped and painted after than time. This shift further increased the investment in the production of white wares relative to gray utility wares.
Figure 18. Plot of the relative frequency of bowl sherds against age for dated pottery assemblages from the Mesa Verde region.
A more direct way to circumvent the closed array problem in comparing the percentages of different pottery wares is to examine accumulation rates. Ideally, one calculates accumulation rates by dividing the amount of pottery in a particular deposit by the amount of time it took for that deposit to form. To do this with any reliability requires extremely well dated assemblages. Unfortunately, very few assemblages can be dated with the degree of accuracy and precision need to calculate reliable accumulation rates. Another way to track accumulation rates of pottery wares is to compare their abundance to the abundance of some other nonpottery artifact class using ratios. By calculating ratios of the abundance of gray and white wares to another artifact class, fluctuations in the abundance of the two pottery wares can be documented independently of one another. The artifact class used to calculate the ratios should be fairly abundant so that sample size effects do not cause wild fluctuations in the ratios from different assemblages. The only artifact class besides in the Mesa Verde region that occurs in great abundance in most assemblages is stone chipping debris, or debitage. Although debitage abundance may vary through time and among contemporary assemblages, the amount of debitage serves as a fixed point of comparison for each assemblage.
Figure 19. Diagram of common vessel forms in utility ware assemblages from the Mesa Verde region.
Figure 20 shows the ratios of white ware and gray utility ware pottery to debitage in assemblages from the Mesa Verde region, which have been grouped into 100 year time intervals. The measure of pottery abundance in these ratios is the weight of sherds in each ware category, while I use counts of flakes and angular debris to quantify the debitage. I used the weight of pottery rather than the count because many of the assemblages vary considerably, and, in some cases, systematically in the degree of fragmentation of the pottery. For example, sherds in plain and neck-banded gray ware assemblages are more than twice as large, on average, as sherds in assemblages dominated by indented corrugation (Figure 21). Such differences in sherd size introduce a systematic bias into the data when counts are used to quantify pottery abundance. The data used in these figures was drawn from reports on single component assemblages from which large samples were recovered by excavation and surface collection using modern methods including probabilistic sampling and screening. Assemblages with large numbers of complete or restorable vessels were avoided. Unfortunately, few reports provide data on the weight of pottery than the count. Consequently, data from only a small fraction of the assemblages meeting these criteria could be used, and no data could be obtained for assemblages dating to the tenth century.
Figure 20. Box plots showing the change through time in estimates of the accumulation rates of utility ware (upper graph) and white ware (lower graph) pottery in assemblages from the Mesa Verde region.
Figure 21. Box plots of the average weight of plain/neck-banded and corrugated utility wares compiled from published sources on pottery assemblages in the Mesa Verde region.
Despite the small number of assemblages, both graphs in Figure 20 show consistent results. The upper graph in Figure 20 demonstrates that the accumulation rates of gray utility pottery were relatively high during the eighth and ninth centuries, and then decreased dramatically after the adoption of full-body corrugation in the early eleventh century. This decrease in accumulation of utility wares may be due primarily to the increased durability of full-body corrugated vessels when used for cooking, an issue I address in more detail in Chapter 7. The lower graph in Figure 20 shows that white ware accumulation rates were very low until after the adoption of full-body indented corrugation, when accumulation gradually increased. Taking these accumulation rate, vessel morphology and ware frequency data together, there is compelling evidence for a decreased investment in utility ware production and an increased investment in white ware production occurring roughly coincident with the adoption of full-body corrugation in the Mesa Verde region.
Utility Vessel Use
Utility ware ceramics in the Southwest are usually thought to have been used primarily for cooking and dry storage. However, none of the defining criteria for this class of pottery, lack of slip, polish, or paint, directly relates to vessel use. Consequently, one cannot assume that by focusing on utility wares, we have adequately identified, or significantly narrowed the range of uses of vessels in this class. Documenting the kinds and intensity of use of vessels in the utility ware class serves two important purposes for this study. First, the use environment is likely to be a significant context of interaction for plain, neck-banded, and full-body corrugated vessels. If the use of utility vessels changed through time, then the vessel-environment interactions also probably changed in ways that could have altered the fitness relationships among the traits of vessels, including various forms of corrugation. Second, designing the studies that document the engineering fitness (cost and performance) of pottery vessels with different combinations of traits, discussed in Chapter 8, requires some knowledge of the use environment in which the various kinds of vessels performed.
In this section, I present data and inferences regarding the kinds and intensity of use of gray utility wares in the Mesa Verde region during the period of change from plain to fully corrugated vessels. Both the morphology (shape and size) and use-related alterations (wear and accumulations) displayed by ancient utility wares provide clues to the use environment of the vessels ( Skibo 1992; Smith 1985). To document variation and change in the use environment of Mesa Verde region utility wares, I recorded data on vessel morphology and use-related alterations in the six gray ware assemblages, and synthesize published information from a variety of sources. The data on vessel morphology and use alterations are presented in separate sections below. However, combined observations of use-related alterations and vessel form offer the strongest sources of data on vessel use. Unfortunately, use alterations are frequently restricted to the lower portions of vessels, and when sherds are examined, it is difficult to identify lower body fragments to a particular vessel form category. To examine form-alteration combinations, I rely on observations made by other investigators on whole or restored vessels, particularly those recovered and analyzed during the Dolores Archaeological Program ( Blinman 1986, 1988a, 1988b).
Vessel Morphology
Archaeologists, as well as producers and users of pottery, frequently equate particular vessel forms with specific uses or functions (e.g., Braun 1980; Henrickson and McDonald 1983; Rice 1987; Shapiro 1984; Smith 1985, 1988). Although the formal design of vessels can certainly affect performance during use ( Braun 1980, 1983; Ericson et al. 1972), in most cases the relationship between form and use is not a necessary one creating the potential for variation and change in use within form categories ( DeBoer and Lathrap 1979; Feathers 1990:62-66; Longacre 1985; Miller 1985; Nelson 1991; Stark 1995b). However, changes through time in the relative abundance of pottery in different form categories can reflect the intensity of use of those forms and possibly particular uses as well.While analyzing the pottery sherds in the six gray ware assemblages from the Mesa Verde region, I recorded two sets of observations that provide information on vessel morphology. For every sherd, I tried to identify the form or shape class of the vessel from which the sherd derived. I also measured rim sherds to estimate the radius of vessel openings, which is correlated with the vessel size, and to quantify the abundance of different vessel shapes and sizes.
Gray ware pottery in the Mesa Verde region is known to encompass a wide range of vessel forms including a variety of jar forms, bowls, pitchers, dippers, and various effigy vessels ( Blinman 1988a, 1988b). However, only five specific vessel forms were identified in the six assemblages I analyzed-wide-mouth jar, olla, seed jar, bowl, and a small cup form (see Figure 19 for illustrations). Wide-mouth jars include necked vessels in which the radius of the opening is at least half the maximum radius of the vessel. Ollas are necked vessels in which the radius of the opening is less than half the maximum radius of the vessel. Seed jars are a closed, jar form, but lacking shoulders and necks. Bowls are open forms in which the radius of the opening is roughly equal to the maximum radius of the vessel. The small cup form was represented by only one piece, and consisted of a crudely fashioned vessel with vertical wall and a flat base. Usually, neck and rim pieces were needed to assign a sherd to one of these specific form categories. Even with neck and rim pieces, distinguishing between wide-mouth jars and ollas could be challenging, particularly in pieces from smaller vessels. Consequently, I employed two more general form categories, indeterminate jar and indeterminate form, in my analysis of the six gray ware assemblages. Frequently, body sherds could be identified as coming from jars based on a smoother exterior surface relative the interior surface, even though the specific jar form could not be identified. These pieces were placed into the indeterminate jar category. More ambiguous pieces were identified as indeterminate form. Most body sherds were assigned to one or the other of these two general form categories.
Estimating the radius of vessel rims in the six assemblages involved measuring the rim curvature and matching the curve to a template of circles with known radii. I measured the curvature by pressing a carpenter's gauge along the rim on the inside surface of the sherd to get an impression of the rim in the gauge. I then matched this impression to one of a set of concentric arcs marked off at one-centimeter radius intervals from one to 26 centimeters. Matching attempts began with small radius arcs and moved out to larger radii until a suitable match was found. If a suitable curve match could not be found, I did not record a radius estimate for that specimen. I also measured the degrees of arc covered by each rim for which a radius estimate was made, and the cord length of all rim sherd arcs to document the size of the rim segment.
Vessel Shape
Most vessel fragments provide very limited information on the shape of the complete vessel from which they are derived. This is certainly true for the collections analyzed in this study. Of the 4174 sherds analyzed, 88.4 percent were identified as indeterminate jar or indeterminate form. Of the sherds identified to one of the six specific form categories, 86.4 percent are rim (43.2 percent) or neck (43.2 percent) fragments where most of the features diagnostic of vessel form occur. Consequently, I compare the abundance of vessel forms in the six assemblages in two ways. First, I compare the weight of all sherds identified to a specific form category regardless of the part of the vessel represented by the sherd. Second, I compare vessel forms using rim sherds only and quantify the abundance based on the degrees of arc covered by the portion of rim present on each sherd. I use these two approaches to quantifying the abundance of different vessel forms because they are likely to introduce distinct biases that, taken together, may give us a clearer picture of vessel form changes ( Pierce and Varien 1999).
Tables 10 and 11 show the abundance of the five different gray ware vessel forms in the six assemblages using sherd weight and degrees of arc, respectively, as the quantifier. Table 12 presents the residuals produced by subtracting the percentages in Table 10 from the percentages in Table 11. It is clear from Table 12 that the rim arc method documented fewer wide-mouth jars relative to other vessel forms than the sherd weight method. There are two likely causes of these differences. First, I was able to identify more non-rim pieces to the wide-mouth jar and olla categories than the other forms resulting in greater weight of sherds than other vessel forms. Second, average olla rims tend to weigh less, but cover more degrees of arc than average wide-mouth jar rims because the olla rim openings are smaller. Given these causes, percentage values between those in Tables 10 and 11 are probably closer to the actual abundance of different vessel forms in these assemblages. However, small samples of sherds with features diagnostic of vessel form may reduce the accuracy and precision of these abundance data in the assemblages with multiple vessel forms represented (see Appendix A).
| Assemblages | Vessel Form Classes | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Wide-Mouth Jar | Olla | Bowl | Seed Jar | Small Cup | Totals | ||||||
| g | % | g | % | g | % | g | % | g | % | g | |
| 5MT2193 | 173 | 56.3 | 34 | 11.0 | 100 | 32.7 | - | - | - | - | 307 |
| 5MT4671 | 497 | 63.0 | 66 | 8.4 | 213 | 27.0 | - | - | 14 | 1.7 | 789 |
| 5MT3868 | 804 | 65.9 | 325 | 26.6 | 86 | 7.1 | 6 | 0.5 | - | - | 1221 |
| 5MT8371 | 475 | 98.8 | 6 | 1.2 | - | - | - | - | - | - | 481 |
| 5MT1786 | 320 | 100 | - | - | - | - | - | - | - | - | 320 |
| 5MT8827 | 754 | 100 | - | - | - | - | - | - | - | - | 754 |
| Totals | 3023 | 78.1 | 460 | 11.1 | 399 | 10.3 | 6 | 0.1 | 14 | 0.3 | 3872 |
Despite the differences between quantification techniques and possible biases from small sample sizes, Tables 10 and 11 show very similar overall trends in the abundance of utility ware vessel forms. Although wide-mouth jars dominate all assemblages, the earliest three assemblages also contain notable amounts of bowls and ollas. Gray ware bowls are most abundant in the earliest assemblage and decrease dramatically, dropping out entirely by the late ninth century AD. Gray ware ollas are most abundant during the late ninth century, but continue to persist into the tenth century. By the tenth century, most bowls and ollas were made as white wares by adding slip and polish, and the gray utility ware pottery in the Mesa Verde region came to be focused almost exclusively on wide-mouth jars. Wide-mouth jars jump from being between one half and two thirds of gray ware vessels in earlier assemblages to around 90 percent gray ware vessels in the tenth century. Blinman (1988b:179) found a similar pattern of change in the relative abundance of gray ware vessel forms in a much larger sample of sherds from trash deposits in the Dolores River area, which suggests that the data presented here are probably accurate.
| Assemblages | Vessel Form Classes | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Wide-Mouth Jar | Olla | Bowl | Seed Jar | Small Cup | Totals | ||||||
| deg. | % | deg. | % | deg. | % | deg. | % | deg. | % | deg. | |
| 5MT2193 | 179 | 54.1 | 17 | 5.1 | 135 | 40.8 | - | - | - | - | 331 |
| 5MT4671 | 555 | 49.7 | 177 | 15.8 | 347 | 31.1 | - | - | 38 | 3.4 | 1117 |
| 5MT3868 | 735 | 55.8 | 490 | 37.2 | 67 | 5.1 | 25 | 1.9 | - | - | 1317 |
| 5MT8371 | 182 | 87.9 | 25 | 12.1 | - | - | - | - | - | - | 207 |
| 5MT1786 | 552 | 100 | - | - | - | - | - | - | - | - | 552 |
| 5MT8827 | 378 | 100 | - | - | - | - | - | - | - | - | 378 |
| Totals | 2581 | 66.1 | 709 | 18.2 | 549 | 14.1 | 25 | 0.6 | 38 | 1.0 | 3902 |
| Assemblages | Vessel Form Classes | ||||
|---|---|---|---|---|---|
| Wide-Mouth Jar | Olla | Bowl | Seed Jar | Small Cup | |
| 5MT2193 | -2.2 | -5.9 | 8.1 | - | - |
| 5MT4671 | -13.3 | 7.4 | 4.1 | - | 1.7 |
| 5MT3868 | -10.1 | 10.6 | -2.0 | 1.4 | - |
| 5MT8371 | -10.9 | 10.9 | - | - | - |
| 5MT1786 | 0.0 | - | - | - | - |
| 5MT8827 | 0.0 | - | - | - | - |
| Totals | -12.0 | 7.1 | 3.8 | 0.5 | 0.7 |
Archaeologists working in the American Southwest commonly refer to wide-mouth jars as cooking pots because of their suitable form and the frequent occurrence of soot on the exterior surface ( Blinman 1988b; Mills 1989). Although some corrugated wide-mouth jars have been found in contexts that suggest use for storage (such as buried in the floors of structures), most of these vessels have sooted exteriors suggesting an earlier use in cooking ( Mills 1989:154-155; Renken 1993; Rohn 1971). If the cooking use inference for wide-mouth jars is correct, it appears that the primary use of gray ware vessels became more specialized through time, and, by the late ninth century, involved cooking almost exclusively.
Other studies of the relative abundance of all vessel forms in the Mesa Verde region, regardless of pottery ware category, have documented that wide-mouth jars were relatively rare in assemblages dating to the seventh century, but their abundance increased during the eighth, ninth, and early tenth centuries ( Blinman 1988b). The occurrence of soot on vessel exteriors indicates that a wider variety of vessel forms, including mainly seed jars, were used for cooking during the seventh century whereas in later assemblages use-related soot appears on wide-mouth jars exclusively ( Blinman 1988a, 1988b; Morris 1939:244-245). A slight decline in the relative abundance of wide-mouth jars during the tenth century coincides with a gradual increase in the abundance of bowls (Figure 18). Some of this change in the relative proportion of vessel forms may derive from a reduction in the accumulation rate of gray wares after the tenth century as discussed earlier in this Chapter. However, change in the bowl to gray ware jar ratio probably also resulted from an increase in the use and discard of bowls. Blinman (1986, 1988b:199) argues that the increase in wide-mouth jars during the eighth, ninth and tenth centuries, as well as other changes, resulted from intensified use of moist cooking or boiling as a part of food preparation. The expanded use of bowls in the tenth and eleventh centuries may also be related to the increased consumption of food prepared by moist cooking.
Vessel Size
For certain vessel forms, particularly wide-mouth jars and bowls, a strong linear correlation exists between the radius at the rim of the vessel and other attributes of vessel size such as volume, height, and weight ( Blinman 1988b; Mills 1989). Given this correlation, it is possible to obtain a measure of vessel size from rim fragments by estimating the radius of the arc of the rim. Error in radius estimates can arise from rims that deviate from circular arcs and small rim segments that make curve matching difficult. In measuring the precision of the curve matching approach to rim radius estimation, Nicklaw (1995:52) found that variation among multiple, independent estimates of the same sherd by the same analyst increased precipitously for rim segments less than 40 mm in length and with an average estimated radius greater than 5 cm. By comparing estimates from rim segments to the true value of the complete or restored vessel, the accuracy of rim radius estimates can be assessed. Although this has not been done for the curve matching technique, Feathers (1990) assessed the accuracy of a different technique and found that estimates decreased in accuracy for rims less than 40 mm long, but were quite accurate for larger rims segments.
Based on these results, Feathers (1990) suggests avoiding radius data from rims less than 40 mm long unless these data do not differ significantly from data derived from larger rims. Because most (86%) of the rim segments analyzed from the six gray ware assemblages are less than 40 mm in length, eliminating the smaller rims would have a devastating effect on sample size. Consequently, a thorough examination of radius results across different rim sizes is called for. Figure 22 shows the relationship between rim arc cord length and the estimated rim radius for all rim sherds in the six assemblages on which I was able to make these measurements. Although the points are quite scattered, a correlation analysis shows a very weak, yet significantly positive (r=.180, p=.002, n=284), relationship. This relationship can arise in two ways. First, smaller rims could result in underestimates of the rim radius given the method of measurement. This is not the case for these data because the small rims account for many of the large radius estimates. Secondly, larger vessels have a tendency to produce larger fragments when they break, and thus, larger vessels will have a tendency to produce rim sherds with greater cord lengths. The increase in the minimum estimated radius with increased rim cord length visible in Figure 22 suggests that the latter cause is responsible for most of the weak correlation seen here. The large number of points and the presence of the single point at the far right side of the graph also contribute to the significant correlation. Fortunately, this relationship between sherd size and vessel size does not bias our estimates of vessel size even if assemblages differ significantly in rim cord length.
Figure 22. Scatter plot of rim radius estimates and rim cord length measurements for all rim sherds in the six utility ware assemblages from the Mesa Verde region.
Given that corrugation was developed and most frequently employed on wide-mouth jars, my analysis of vessel size will focus on that form category. Figure 23 shows box and whiskers plots of rim radius weighted by degrees of arc and rim cord length for sherds identified as wide-mouth jars and indeterminate jars in the six gray ware assemblages. I include indeterminate jar rims in this analysis of vessel size to increase the sample size and because most of the rims identified to this form category are very likely to be from wide-mouth jars, especially in the later assemblages. The cord length plot shows that the earliest and latest assemblages have the longest rims while the assemblages in between are generally more fragmented. In contrast, the plot of rim radius shows a consistent trend toward larger jars through time. The differences in temporal tends between rim cord length and rim radius further indicate that variation in rim size among assemblages has not compromised the reliability of the rim radius data to document changes in vessel size. However, small sample size may reduce the accuracy and precision of these vessel size data (see Appendix A).
Figure 23. Box plots of rim cord length (upper graph, N=count) and rim radius (lower graph, N=degrees of arc) for wide-mouth and indeterminate jar forms in the six utility ware assemblages from the Mesa Verde region.
Cumulative percentage graphs of rim radii weighted by degrees of arc for the six assemblages (Figure 24) also show an increase in the abundance of larger wide-mouth jars through time while small jars also continued to be made. In addition, sharp inflections in the graph of the latest assemblage (5MT8827) show the differentiation of rim radii into three fairly discrete size modes. Both the general trend toward larger gray ware jars and the occurrence of size modes have been documented in a much larger sample of sherds and complete vessels from the Mesa Verde region ( Blinman 1988b; Mills 1989) and elsewhere in the northern Southwest ( Mills 1993:344). Thus, these patterns are probably robust and not a reflection of measurement error or sample size bias. What has not been noted previously is that multiple discrete size modes may have appeared for the first time in the eleventh century in the Mesa Verde region coincident with the adoption of corrugation. A similar timing of the appearance of wide-mouth jar size modes occurs in data on radius estimates of rim sherds recovered from trash deposits excavated and analyzed by the Dolores Archaeological Program ( Blinman 1988b: 184). However, data on the actual volume of a much smaller, though still substantial, sample of complete vessels recovered and restored during the same project conflict with the data from rim sherds and suggest that discrete wide-mouth jar size modes appeared earlier than the eleventh century ( Blinman 1988b:128).
Figure 24. Cumulative percentage graphs of estimated rim radii weighted by degrees of arc for wide-mouth and indeterminate jars in the six utility ware assemblages from the Mesa Verde region.
Both the increase in the largest size of wide-mouth jars and the development of discrete size modes suggest that there may have been some changes through time in the way these vessels were used. If wide-mouth jars were used mainly as cooking pots, the addition of larger and larger vessels through may signal an increase in the size of groups involved in food preparation or consumption. If wide-mouth jars were also used in the storing of dry goods, larger vessels may also indicate increased storage needs. The development of discrete size modes may be linked to more formalized production or use of wide-mouth jars. If size modes appeared for the first time in the eleventh century, the correlation between the adoption of full-body corrugation and the appearance of more formal production or use contexts may be significant.
Use Alterations
In the analysis of the six gray ware assemblages I examined for this study, I recorded data on two kinds of use-related alterations for each surface of all sherds-the type and degree of wear and the presence of accumulations. Wear types I monitored include abrasion, pitting, scratches, polish, flaking, and spalling, and the degree of wear was described as light, moderate, or heavy. The surface accumulations recorded include soot, charred material, uncharred material, and stains. These kinds of use alterations can be difficult to distinguish from traces left during the manufacture process and alterations produced after the sherds were discarded. Although I was extremely conservative in identifying traces as use-related alterations, it is important in analyzing these data to consider other possible sources for the ware and accumulations.
Table 13 shows the amount of pottery identified to different attritional wear classes. Most of the sherds show no clear signs of wear, which is common in pottery assemblages. However, small yet notable amounts of pottery are spalled and pitted. Pitting refers to clusters of small, shallow indentations or divots in the surface of the pottery, while spalled sherds include those with partially or completely missing surfaces that cannot be attributed to abrasion or pitting. Both of these wear classes could be produced through use or by post-depositional alteration. However, the tendency for spalling and pitting to occur preferentially on exterior and interior surfaces respectively, and for both types of wear to appear mainly on basal fragments (Figure 25) indicates that they were formed primarily through use. Use-life experiments presented in Chapter 7 also indicate that basal pitting and spalling are produced on some vessels as a consequence of their regular use in cooking. That many gray ware vessels were used for cooking is suggested by the fact that over 40 percent of all base fragments in the six assemblages show some pitting of the interior surface.
| Wear Class | Surface | Assemblages | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 5MT2193 | 5MT4671 | 5MT3868 | 5MT8371 | 5MT1786 | 5MT8827 | ||||||||
| g | % | g | % | g | % | g | % | g | % | g | % | ||
| No Wear | Interior | 2067 | 89.7 | 5194 | 86.1 | 6099 | 84.0 | 1245 | 80.0 | 2770 | 83.3 | 3235 | 92.6 |
| Exterior | 2230 | 96.7 | 5684 | 94.0 | 6781 | 93.1 | 1331 | 85.0 | 3247 | 97.5 | 3343 | 95.7 | |
| Spalling | Interior | 16 | 0.7 | 114 | 1.9 | 144 | 1.9 | 57 | 3.7 | 65 | 2.0 | 1 | 0.0 |
| Exterior | 31 | 1.3 | 202 | 3.3 | 223 | 3.1 | 212 | 13.6 | 42 | 1.3 | 0 | 0.0 | |
| Pitting | Interior | 201 | 8.7 | 636 | 10.5 | 942 | 13.0 | 241 | 15.5 | 447 | 13.5 | 147 | 4.2 |
| Exterior | 1 | 0.1 | 7 | 0.1 | 53 | 0.7 | 0 | 0.0 | 9 | 0.3 | 0 | 0.0 | |
| Abrasion | Interior | 19 | 0.8 | 15 | 0.9 | 58 | 0.8 | 13 | 0.8 | 40 | 1.2 | 2 | 0.1 |
| Exterior | 40 | 1.7 | 153 | 2.5 | 154 | 2.1 | 22 | 1.4 | 33 | 1.0 | 31 | 0.9 | |
| Other | Interior | 3 | 0.1 | 35 | 0.6 | 15 | 0.2 | 0 | 0.0 | 3 | 0.1 | 108 | 3.1 |
| Exterior | 3 | 0.1 | 0 | 0.0 | 74 | 1.0 | 0 | 0.0 | 0 | 0.0 | 118 | 3.4 | |
| Percentages calculated separately for each surface. | |||||||||||||
Figure 25. Relative abundance of spalling and interior surface pitting of different vessel parts in data combined from all six of the utility ware assemblages from the Mesa Verde region. N = grams of pottery.
The incidences of spalling and interior surface pitting, as measured by the percentage of pottery with these alterations in each assemblage, show the same trajectory in time (Figure 26). Both of these forms of use-alteration increase from relatively low levels in the eighth century AD to a peak in the tenth century and fall off rapidly in the eleventh century. If the changes in the incidence of spalling and pitting in plain surface pottery are related to the intensity of use in cooking, then the wear pattern data correspond with the vessel form data in indicating intensified use of gray wares in cooking through the ninth and tenth centuries. However, the reduction in the range of vessel forms, and presumably uses, among gray wares could also account for the increased incidence of pitting and spalling in the tenth century with a greater proportion of vessels being used in cooking rather a more intense use of individual vessels. Regardless of whether it was the intensity of use of individual vessels or the proportion of vessels used for cooking, the wear data indicate that use of gray wares for cooking increased from the eighth to the tenth centuries. The subsequent decrease in spalling and pitting after the adoption of full-body corrugation probably stems from the effects of corrugation on the nature of thermal stresses in the vessel wall (see the discussion in Chapter 7) rather than indicating a reduction in cooking intensity.
Figure 26. Relative abundance of spalling and interior surface pitting in the six utility ware assemblages from the Mesa Verde region. Data labels are grams of pottery and error bars represent the likely temporal span of each assemblage.
Use-related accumulations are fairly rare in the six assemblages except for soot adhering to the exterior surface of sherds, which occurs in all six assemblages to varying degrees (Table 14). Soot accumulations are quite abundant in some assemblages while almost entire lacking in others. Soot adheres to vessel surfaces when the vessels rest for long periods over a source of smoke. Although pottery vessels can be exposed to smoke in various ways, the most common context in which this occurs is most likely through their use in cooking over an open fire. When vessels are used over an open fire, soot tends to accumulate in patterned ways. Experiments and ethnographic observations conducted by Jim Skibo (1992) indicate that soot accumulates most readily in the widest portions of the vessel, which have the greatest exposure to smoke rising from the fire. If the base of the vessel is suspended above the fire soot will also accumulate on the base. However, if the bottom of the vessel is left in the flames or resting on burning embers soot does not accumulate on the base. I have seen both patterns of soot accumulation in my experience with whole and restored wide-mouth utility ware vessels from the Mesa Verde region. Almost all (99%) of the exterior sooting in the six utility ware assemblages occurs on wide-mouth and indeterminate jar fragments. The exceptions include a single olla sherd from Dos Casas Hamlet (5MT2193) and a single bowl sherd from Periman Hamlet (5MT4671), the two earliest assemblages. The distribution of soot across different vessel parts in all six of the utility ware assemblages (Figure 27) also suggests that both patterns of soot accumulation were present as well. The occurrence of these soot accumulation patterns on plain, neck-banded, and corrugated gray ware jars indicates that many of these vessels were used in cooking.
| Residue | Surface | Assemblages | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 5MT2193 | 5MT4671 | 5MT3868 | 5MT8371 | 5MT1786 | 5MT8827 | ||||||||
| g | % | g | % | g | % | g | % | g | % | g | % | ||
| None | Interior | 1713 | 74.3 | 5525 | 91.8 | 6760 | 93.3 | 1483 | 95.1 | 3298 | 99.1 | 2907 | 83.2 |
| Exterior | 1287 | 55.8 | 4293 | 71.1 | 5814 | 79.8 | 1170 | 74.8 | 3142 | 94.5 | 1873 | 53.6 | |
| Soot | Interior | 319 | 13.8 | 149 | 2.5 | 24 | 0.3 | 0 | 0.0 | 2 | 0.1 | 89 | 2.6 |
| Exterior | 918 | 39.8 | 1748 | 28.9 | 1418 | 19.5 | 395 | 25.2 | 183 | 5.5 | 1593 | 45.6 | |
| Stain | Interior | 143 | 6.2 | 60 | 1.0 | 165 | 2.3 | 70 | 4.5 | 0 | 0.0 | 347 | 9.9 |
| Exterior | 100 | 4.3 | 0 | 0.0 | 57 | 0.8 | 0 | 0.0 | 0 | 0.0 | 26 | 0.7 | |
| Charred Material |
Interior | 131 | 5.7 | 284 | 4.7 | 278 | 3.8 | 0 | 0.0 | 16 | 0.5 | 149 | 4.3 |
| Exterior | 0 | 0.0 | 0 | 0.0 | 0 | 0.0 | 0 | 0.0 | 0 | 0.0 | 0 | 0.0 | |
| Other | Interior | 0 | 0.0 | 0 | 0.0 | 16 | 0.2 | 7 | 0.4 | 13 | 0.4 | 0 | 0.0 |
| Exterior | 0 | 0.0 | 0 | 0.0 | 0 | 0.0 | 0 | 0.0 | 0 | 0.0 | 0 | 0.0 | |
| Percentages calculated separately for each surface. | |||||||||||||
Figure 27. Relative abundance of sherds with surface soot accumulations for different vessel parts in combined data from all six of the utility ware assemblages from the Mesa Verde region. Data labels are grams of pottery.
The incidence of soot accumulations among the six assemblages shows a decline through time until the middle of the eleventh century AD when it appears to increase rapidly (Table 14). This is not what I would expect if the intensity of use in cooking was increasing through time. One possible explanation is that differences in sherd size among assemblages have biased the data on soot accumulations because the variation in sooting is very similar to that in sherd size (see Appendix A). However there is strong evidence that this pattern is due mainly to post-depositional alteration rather than variation in use or sherd size. The three assemblages with the lowest amount of exterior sooting, Duckfoot (5MT3868), 5MT8371, and Gnatsville (5MT1786), are also the only assemblages for which the entire sample comes from the surface or from relatively shallow midden deposits. Comparison of the amount of sooting on sherds from shallow deposits to those from deeply buried deposits in pit features and structures shows considerable differences. At Dos Casas Hamlet (5MT2193), 2 percent of the sherds from the surface midden are sooted while 51 percent of the sherds from the trash fill of Pit Structure 1 show sooting. A similar pattern is found a Periman Hamlet (5MT4671) at which 14 percent of the surface midden sherds are sooted and 38 percent of the sherds from the fill of pit features show sooting. These results indicate that exposure to the intense weathering conditions at or near the surface results in the erosion of accumulated soot from the surface of pottery.
One possible way to control for this post-depositional bias is to only compare soot accumulations in assemblages from protected depositional contexts. Fortunately, the three protected contexts occur at opposite ends of the time range covered by the six assemblages. Fifty-one and 38 percent of the gray ware pottery from protected contexts at Dos Casas and Periman Hamlets have soot accumulations, while 46% of the pottery from Dobbins Stockade (5MT8827) have soot of their exterior surfaces. These data indicate that the degree of sooting was probably fairly consistent across the period covered by these assemblages, suggesting that cooking may have been a common use for gray ware jars throughout the period during which corrugation was developed.
Summary of Use Evidence
Both the vessel morphology and use alteration data indicate that the use of utility ware vessels became increasingly focused on cooking during the ninth and tenth centuries in the Mesa Verde region. In early assemblages, pottery in the gray utility ware category was made into a wide variety of vessel forms and shows little cooking-related ware. Through time, the variety of vessel forms decreased to a single form, wide-mouth jars, the size of these jars increased and became more formalized, and the relative abundance of cooking-related ware increased. Although soot accumulations are as common in earliest assemblages as they are in later assemblages, soot occurs on a wider variety of vessel forms in these earliest assemblages. These data indicate that cooking was always part of the use environment of utility wares, but utility vessels in the Mesa Verde region became increasingly specialized in cooking through time. By the eleventh century when full-body corrugation was adopted, cooking appears to have been the primary use environment of utility wares.
Other data from the Mesa Verde region indicate that an increased production and use of agricultural products occurred simultaneously with the greater specialization of utility ware use during the ninth and tenth centuries. These data include greater investment in larger and more formalized storage facilities ( Gross 1986), increased abundance of two-hand manos used for grinding grain on slab metates ( Phagan 1986), and a reduction in the ratio of animal bones to cooking pot fragments in very large samples compiled from several sites suggesting that cooking became more focused on plant materials through time ( Blinman 1986). These changes also coincide the introduction of a new corn variety, Mais de Ocho, which produces higher yields and has a softer endosperm more suitable for milling than the earlier flint corn varieties ( Adams 1994; Cutler and Meyer 1965; Galinat 1985; Galinat and Gunnerson 1963; Galinat et al. 1970, Mathews 1986).
If corn grains were increasingly processed into meal, as the increase in two-hand manos and the use of Mais de Ocho suggest, then the cooking of corn products would also have shifted from roasting of whole ears to other cooking techniques more suitable for meal. Corn meal can be cooked in two ways: (1) boiling with water and, potentially, other food stuffs into a mush or stew, and (2) baking or grilling into bread or tortillas. The lack of griddle stones and other artifacts and features associated with baking and grilling until the fifteenth century in the northern Southwest suggests that corn meal was cooked primarily by moist cooking or boiling until very late in the prehistoric period ( Snow 1990). Thus, cooking in pots increased in importance during the ninth and tenth centuries in the Mesa Verde region at the same time that there is increased evidence for the use of utility ware vessels in cooking.
Although detailed evidence is lacking, cooking appears to have been the primary use of Puebloan utility wares for the next 400 years. However, apparent changes in the processing of corn during the fifteenth century suggest that the importance of cooking corn meal in pots may have decreased at the same time as the shift back to manufacture of plain utility wares. During the fifteenth century in the Rio Grande region and possibly elsewhere, a suite of traits used in bread or tortilla making, including griddle stones or comales, rectangular hearths, and fire dogs (stones used to support the comales over the hearth), appeared for the first time ( Snow 1990). Presumably introduced from Mexico where this technology has a much longer history, this bread or tortilla making probably also involved soaking or mixing the corn with lime or alkali which significantly enhances its nutritional value ( Katz et al. 1974). This adoption of a flat bread technology probably resulted in much less corn being cooked into mush and gruel. In fact, a general lack of information on porridge cooking in historical and ethnobotanical information on the Pueblos may indicate that this method of cooking corn was of little significance in historic times.
Return to Table of Contents
Go to Chapter 7
