How to Make an Unfired Clay Cooking Pot - Springer Link

J Archaeol Method Theory (2009) 16:33–50 DOI 10.1007/s10816-009-9061-4

How to Make an Unfired Clay Cooking Pot: Understanding the Technological Choices Made by Arctic Potters Karen G. Harry & Lisa Frink & Brendan O’Toole & Andreas Charest

Published online: 10 February 2009 # Springer Science + Business Media, LLC 2009

Abstract Between about 500 A.D. and the late nineteenth century, clay cooking pots associated with the Thule culture were produced in the Arctic region. Ethnographic and archaeological records indicate that these vessels were typically underfired (often even unfired), highly porous, and easily broken. Despite these characteristics, the evidence indicates that they were used to heat water over open fires. In this paper, we examine how Arctic potters were able to produce unsintered vessels capable of holding liquids without disintegrating. We conclude that the application of seal oil and seal blood to the pot’s surface was the key to their success. Keywords Ceramic technology . Arctic . Experimental archaeology . Traditional technologies Clay cooking pots were produced and used in the Arctic from about 2,500 years ago until the middle or late nineteenth century (Fig. 1). Although the earliest vessels tended to be thin-walled and relatively well-fired, by 1,000 A.D. these containers were replaced by ones having thicker walls and coarse-textured, soft pastes. Ceramics recovered from the later periods have a marked tendency to crumble and exfoliate, and many were either underfired or not fired at all. The technology of these later vessels differs in nearly every significant way from that exhibited by the typical clay cooking pot found in other areas of the world. In fact, these Arctic cooking pots break nearly every engineering rule about how a ceramic cooking pot should be constructed (see Frink and Harry 2008) and—at least for the unfired or K. G. Harry (*) : L. Frink : A. Charest Department of Anthropology and Ethnic Studies, University of Nevada, Las Vegas, 4505 Maryland Parkway, P.O. Box 45003, Las Vegas, NV 89154-4003, USA e-mail: [email protected] B. O’Toole Department of Mechanical Engineering, University of Nevada, Las Vegas, 4505 Maryland Parkway, P.O. Box 4005, Las Vegas, NV 89154-4005, USA

34 Fig. 1 Areas discussed in text. Shaded areas indicate regions having pottery in prehistoric and historic periods; hachured and shaded area had pottery prehistorically only.

Harry et al.

Arctic Ocean Siberia

Alaska

Tununak

Bering Sea 0

km

1000

very low-fired ones—seemingly should never have been able to “work” as cooking vessels. Arctic vessels are not alone in contradicting our expectations about how cooking pots should be constructed. Ceramic engineering principles suggest that vessels best designed to transfer heat and withstand the stresses resulting from repeated cooling and heating cycles are those that are thin-walled, globular in shape, and have a moderate-to-low porosity. Although most cooking pots associated with sedentary societies do have these characteristics, many such containers from mobile hunter– gatherer societies do not. This mismatch between our expectations and the attributes of hunter–gatherer pottery suggests that the performance characteristics most important to hunter–gatherers differed from those most important to sedentary agriculturalists (Frink and Harry 2008; Schiffer and Skibo 1987; Skibo et al. 1989). Accordingly, it becomes important to study hunter–gatherer pottery in its own right and to develop models and middle-range theories specific to hunter–gatherer use of ceramic technology. This paper represents one attempt to meet this goal. As such, it joins a growing body of research focusing on ceramic manufacture and use in hunter–gatherer societies (Eerkens 2003, 2004; Eerkens et al. 2002; Frink and Harry 2008; Reid 1989, 1990; Rodrίguez 1995; Roosevelt 1995; Sassman 1992, 1995; Skibo and Blinman 1999). In this paper, we explore how Arctic potters managed to make underfired or even unfired clay containers “work” as cooking vessels. The use of unsintered clay pots for cooking purposes was not limited to the Arctic; such containers were also manufactured and used by other hunters and gatherers in the world, including those living in South Africa (Sampson 1988) and various areas of the United States (Ewers 1945; Kelly 1976; Ray 1932; Sapir 1923; Skibo et al. 1989; Wissler 1910). According to modern ceramic engineering principles, such containers should have performed poorly (if at all) as cooking vessels—unsintered clays, by definition, dissolve when exposed to liquids. Additionally, the high porosity associated with unfired or low-fired clays

How to Make an Unfired Clay Cooking Pot

35

would have substantially reduced the heating capabilities of the pot. To explore how hunter–gatherer potters overcame these problems, we investigate the manufacturing technologies used with one group of underfired cooking containers—that of the Arctic cooking pot. Because no analog exists for this technology today, we found that we could not turn to modern ceramic engineering principles to explain this issue. Instead, to address this question, we adopted a multipronged approach that incorporated information from a variety of sources, including ethnohistoric accounts, indigenous knowledge held today by Arctic peoples, and information derived from experimental archaeology. Our results indicate that Arctic potters were knowledgeable experts of the properties exhibited by various resources found in their environment and that they skillfully manipulated these resources to produce cooking containers that worked well despite not being fully sintered. The results of this study demonstrate the importance of combining indigenous knowledge with experimental data to investigate prehistoric technology, particularly when trying to understand technologies that have no analog in the modern-day world.

Ethnographic and Archaeological Descriptions of Arctic Ceramics Two primary types of pottery are found in the Arctic region: that associated with the Norton tradition (ca. 500 B.C.–500/1,000 A.D.) and that associated with the Thule tradition (ca. 1,000–1,600 A.D.) and its descendant historic Eskimo cultures. The research presented in this study focuses on the latter pottery type.1 Most of what is known about Thule-style pottery comes from the ethnographic record. Because such vessels did not fall out of use until relatively late in the historic period, considerable information about these ceramics can be gleaned from ethnographic accounts. Additional information derives from the examination of archaeological ceramics and historic collections. A comparison of archaeological sherds with historic pottery and ethnographic accounts indicates that no major technological changes in pottery production occurred between the prehistoric Thule and historic periods. Therefore, we propose that both archaeological and historic data can be used to inform on how Thule-style pottery was produced. In this section, we review data obtained from these two information sources to investigate the pottery production process. Thule-style vessels were invariably flat-bottomed and flower pot (slightly flaring walls)-shaped or bucket (straight-walled)-shaped (Fig. 2). Both organic and inorganic materials are reported to have been used as tempering agents. Inorganic materials added to the raw clays included coarse pebbles (de Laguna 1939:334), sand (de Laguna 2000:75; Gordon 1906:83; Nelson 1899:201), and crushed rock Pottery associated with the Norton tradition is typically described as being “relatively thin-walled, well fired, [and] tempered with a mixture of organic fibres and sand” (McGhee 1980:44). Thule pottery, in contrast, is described as being “plain, thick, poorly fired [and] tempered with sand and gravel” (McGhee 1980:45). In practice, the difference between the two types appears to be one of a general trend rather than a sharp distinction, and there is a great deal of variation within each tradition. In this paper, we deal only with the Thule cooking pot. Although we believe that the differences between the two technologies deserve exploration, we find that the issue is beyond our ability to explore at this time given the relative paucity of information pertaining to Norton ceramics and to the environmental and cultural contexts of these cultures. 1

36

Harry et al.

Fig. 2 Arctic cooking pots. Pots were collected from the Kuskokwim River region of Alaska (top left); from near Nome, Alaska (top right); from Dering, Alaska (bottom left), and eastern Siberia (bottom right). All vessels are in the Museum of the North American Indian (MNAI) collection, photographs reproduced courtesy of MNAI.

(Oswalt 1952:21, 1955). Organic materials commonly added were feathers (de Laguna 1947:141–143, 227–230), grass (Collins 1937:167; de Laguna 1947:228– 229; Nelson 1899:201), and animal hair (de Laguna 1947:227). Ethnographic accounts indicate that there was great variation in the paste recipes used; added tempers might consist only of organic materials, only of inorganic ones, or of a mixture of both (see de Laguna 1947:140–149, 226–249). Repeated references also are made to the addition of sea mammal oil (de Laguna 2000:125), fish grease (de Laguna 1947:1412, 2000:119), and/or animal blood (de Laguna 1939:341, 1947:1412; Fienup-Riordan 1975:14; Gordon 1906:83; Spencer 1959:471–472). But descriptions vary as to how and when these materials were added. Whereas some accounts indicate that these substances were mixed into the raw clay during paste preparation (de Laguna 1947:1412, 2000:125; Fienup-Riordan 1975:14; Gordon 1906:83; Spencer 1959:471–472), others report that they were applied to the vessel surfaces before or after firing or after each use of the vessel (Bogoras 1904:186; de Laguna 1939:339, 2000:119, 135). Finally, at least two accounts refer to a process whereby the clay container was filled with oil (or an oily broth) and allowed to stand until the lubricant had permeated the vessel wall (Osgood 1940:146–149; de Laguna 1947:235). Any of a number of methods could be used to form the vessel. Small vessels were often modeled from a single lump of clay which was paddled into shape (de Laguna 1940:64). In other instances, vessels were shaped by paddling clay against a form or mold (de Laguna 2000:141; Zhushchikhovskaya 2005:51–54). The slab construction technique, in which rectangular slabs of clay were added vertically to the edges of a basal slab disc (Oswalt 1952, 1955; Zhuskchikhovskaya and Shubina 2006:101– 102), was often used for larger vessels. Alternatively, pots could be formed by


37

attaching a thick coil of clay to the slab base and then pinching the coil to thin and shape it into vessel walls (Fienup-Riordan 1975:14–15). Finally, some large vessels were formed by patching together small pieces of clay (Osgood 1940:146–149) or by connecting and thinning clay coils (de Laguna 1940:64; Nelson 1899:201). Surface finishing techniques were similarly variable; some surfaces were intentionally roughened while others were lightly smoothed. Although surfaces were never polished, ethnographic and archaeological data indicate that they were sometimes smoothed with fingers, scrapers, or wads of grass (de Laguna 1947:143). At other times, they were incised with wooden or other implements (Gordon 1906:83), or textured using basketry molds or incised paddles on the wet clay (Zhushchikhovskaya 2005:22). On yet other vessels, the surface was coated with a clay–water mixture (Zhushchikhovskaya 2005:75) or smoothed with water (de Laguna 2000:243) to seal the porous walls. Once formed, the pots were dried or fired. Again, however, descriptions of this process vary substantially. Some accounts report that the pots were not fired at all, but were merely hardened in the sun (de Laguna 1947:131, 235, 2000:135, 141; Mathiassen 1927:106). Others refer to a process whereby the pots were placed near the fire to dry, but the descriptions suggest they were not exposed to sufficiently hot temperatures for the sintering process to begin. For example, in describing pottery manufacture in the Kotzebue Sound area, Stefánsson (1914:312) reported that: When shaped the pot was set beside a small fire and slowly dried, being turned a quarter round every little while. A pot would dry between morning and evening…. The pots were never burned, not even allowed to get very hot in drying. A similar process was observed by Osgood (1940:146–149) in the Yukon– Kuskokwim region: When the pot has been shaped it is moved… about 3 to 4 feet from the fire and allowed to dry slowly. This takes about two days, the pot being turned from time to time and tested by tapping with a little stick in order to determine its condition of dryness by the sound. When the wall of the pot is dry, it is tipped over so that the bottom also dries. After this, a little fire is made inside with shavings to burn off the edges of the feathers which roughen the surface of the pot. In contrast, Nelson (1899:228) observed a very different firing process in the Bering Strait area. There, he witnessed vessels being fired by “building a fire both inside and outside” the vessel, after which time the vessels were “baked for an hour or two with as great a heat as can be obtained.” The variation described in drying and firing practices may have been geographically patterned. This possibility was raised by de Laguna (1947:227), who suggested that pottery from the Yukon–Kuskokwim region was better fired than that from areas further north. However, the variation in firing practices may have also resulted from localized or temporal differences in firewood availability or from chance differences in the weather on the day or days that any particular pot was made. As we have discussed elsewhere (Frink and Harry 2008), the cool temperatures and high humidity of the Arctic region would have made it very difficult to fire pottery, since pots that are insufficiently dry are likely to explode during the firing process. Arctic

38

Harry et al.

potters may have found it impossible to adequately dry their newly formed vessels on cool and humid days. On such days, they may have elected to leave their pots either unfired or only slightly fired. On relatively warmer or drier days when it may have been possible to more fully dry their pots, they may have elected to expose their vessels to higher temperatures in order to obtain a harder and more durable product. Additional research is needed to resolve this issue. Attributes of Thule sherds indicate that the production techniques described in the ethnographic accounts were a continuation of those used during prehistory. The pastes are often porous, a result of using organic temper and a low firing process. Inorganic tempering agents are present as well, and the sherds often exhibit thick organic rinds like those produced experimentally when oil or blood is added to a pot’s surface (Harry, K.G., Frink, L., Swink, C., and Dangerfield, C. Ms. on file, Department of Anthropology & Ethnic Studies, University of Nevada Las Vegas, Las Vegas; Stimmell and Stromberg 1986:243). The ceramics also reflect substantial variation in firing technology. Whereas uncarbonized hair and plant fibers have been observed in some Thule sherds, suggesting that they had not been fired highly enough to combust these materials (Stimmell and Stromberg 1986), attributes of other ceramics suggest that they had been exposed to high enough temperatures to at least begin the sintering process (Harry, K.G., Frink, L., Swink, C., and Dangerfield, C. Ms. on file, Department of Anthropology & Ethnic Studies, University of Nevada Las Vegas, Las Vegas). The variation observed in Thule pottery and in the historic accounts is not surprising given the large area over which this pottery was made. (To provide some indication of the size of the region, we note that the area is larger than that encompassed by the states of Arizona and New Mexico combined.) There is as yet relatively little understanding of the nature of this variation. Nonetheless, certain recurrent characteristics are found among the Arctic cooking pots, including the use of organic and inorganic tempers, the addition of animal blood and sea mammal oil to the vessels, and—especially—the presence of soft, porous, and friable pastes that disintegrate easily. Because cooking pots—which are designed to hold and heat liquids—must be relatively impermeable to liquid and able to withstand contact with water, we found the latter characteristics especially perplexing.

The Enigma of the Underfired Cooking Pot Despite the porosity and fragile nature of the Thule and historic vessels, the ethnographic evidence clearly indicates that they were used for cooking purposes. There exist several references to cooking with clay vessels in the historic literature (Ray 1975:117; Spencer 1959:53–54; Stefánsson 1962:176–178), but no references regarding any other uses for these containers. (The only other type of pottery made in the Arctic was the oil lamp, which can be easily distinguished in archaeological sherds from cooking containers by their shapes and thicknesses.) Vessels were called egan, which etymologically is said to closely relate to the concepts of “cook” and “boiled food” (O’Leary 1999:6). Despite the use of the term “boiled” in the above translation, ethnographic accounts indicate that the so-called boiled foods were, in fact, merely briefly immersed in simmering liquids (Frink and Harry 2008; Spray 2002:36).


39

Unless sealed, highly porous walls are not well-suited for cooking activities because they allow the liquid to seep out of the vessel. Not only does this result in a loss of liquid, but it is difficult to boil water in such containers (Schiffer 1990; Skibo 1992). This is because any heat generated within the contents is constantly being dissipated through the release of hot water or steam through the vessel wall. Similarly, we would not expect unsintered vessels to perform adequately as cooking pots since their pastes will disintegrate when exposed to liquids. Although the temperatures needed to solidify clay into a ceramic material (i.e., one that is permanently hardened and will not fall apart in water) will vary depending on such factors as the clay mineralogy and length of firing; in general, temperatures above 600°C or so are required (Gibson and Woods 1997:115; Velde and Druc 1999:53). At temperatures approaching 600°C, the clays will begin to partially harden but may disintegrate when exposed to water for long enough periods of time. Our own experiments suggest that pottery fired below about 400°C will completely fall apart when exposed to water, whereas that fired to between about 400°C and 600°C will retain its general shape but become very soft and will crumble easily with the slightest impact. The ethnographic accounts suggest that temperatures above 600°C were seldom achieved by Arctic potters and that the low temperatures did, as we would expect, result in vessels that easily absorbed water and were prone to disintegration when wet. For example, Osgood (1940:146–149) reports that vessels were stored off the ground to keep them from absorbing moisture from the damp earth and that water would never be stored in pots as it was said that this would soften the clay. Stefánsson (1914:322) similarly relates that the unfired pots he observed “broke easily and spoiled in long spells of wet weather.” Elsewhere (Frink and Harry 2008), we have argued that the production of such weak and porous vessels resulted from the manufacturing techniques necessitated by the cool, wet Arctic weather. In other words, we propose that these were not desirable characteristics, but that they were accepted by Arctic potters as a necessary trade-off to the ability to have pots at all. The use of organic tempers and/or coarse inorganic ones would have created fragile pots, which would have been made even weaker by the practice of not fully firing the vessels. Such production techniques, however, would have avoided the problem of having damp vessels explode during the firing process. Vessels fired while still damp can shatter when steam becomes trapped inside the vessel wall; this is especially likely to occur when the vessel is made of a fine-textured paste. In the cool, damp Arctic weather, it would have been difficult or even impossible for potters to fully dry their pots prior to firing. The highly porous pastes resulting from the use of organic and coarse organic tempers would have minimized the problem by allowing steam to escape more easily. The problem would have been further minimized by the low firing temperatures, which would have kept steam from becoming trapped during the sintering process. The question remains, however, how they were able to make these low-fired, unsintered, and porous containers capable of holding and heating liquids. To address this issue, a number of archaeological experiments were undertaken. These included both informal experiments conducted to try to replicate the manufacture of the pots and more structured trials undertaken with the goal of evaluating the effects of particular technologies on various performance characteristics. Both types of experiments played a role in addressing the research issue. The replication experiments allowed us to identify

40

Harry et al.

and refine the probable production methods and to narrow down the variables to be more formally tested in the laboratory experiments. The latter experiments, in contrast, enabled us to hold as many variables constant as possible so that we could isolate and test the influence of one variable at a time.

Preliminary Replication Experiments Replication experiments were carried out in both field and laboratory settings. The field experiments were undertaken in the summer of 2005 in the village of Tununak, located on Nelson Island in Western Alaska (see Fig. 1). These experiments were carried out with the assistance of two expert potters skilled in hand-building techniques. Among other reasons, Tununak was selected as the location for the field experiments because oral accounts indicate that, historically, clays used for pottery manufacture were collected near that village (Friday 1983). The field experiments were considered integral to our technological studies because they enabled us to make pottery using similar clays and the same types of raw materials (notably, seal oil and seal blood) as were used by prehistoric and historic Arctic potters. Additionally, because the environmental setting and climatic conditions were similar, we could evaluate the difficulties and advantages associated with the use of different manufacturing techniques for these craftspersons. Upon return from the field, additional replication experiments were conducted on the campus of the University of Nevada, Las Vegas. These experiments allowed us to follow-up on questions resulting from the field experiments, although of course under very different environmental conditions. The most relevant of these differences was the low humidity (averaging less than 25% at the time of our experiments) in Las Vegas compared to the very high humidity (averaging between 80% and 90%) experienced in Tununak. The field experiments were carried out using local sources of clay; these included clays collected from a nearby mountain that were identified by several of the village elders as being those used historically for pottery manufacture. Additionally, we used clays collected from a nearby beach; elders identified these sediments as having been used historically to caulk the exterior of kayaks. The mountain clays were found to be extremely plastic and easy to mold. However, they had a severe tendency to crack during the drying process. The severity of the cracking was related to both the rapidity of the drying process and the porosity of the pastes. Thus, they tended to crack more in the arid Las Vegas climate than in the humid Arctic one and more greatly in the Arctic when attempts were made to speed up the drying process (for example, by drying the pots overnight in an electric oven). They also cracked more when tempered entirely with inorganic tempers than when organic ones were used. The beach clays were less plastic but still tended to crack during drying. The results of these various replication experiments are more fully described in Harry, K.G., Frink, L., Swink, C., and Dangerfield, C. (Ms. on file, Department of Anthropology & Ethnic Studies, University of Nevada Las Vegas, Las Vegas). For the purposes of this paper, however, the relevant findings from these experiments have to do with the application of animal blood and seal oil. In particular, we found that it was impossible to work with pastes in which blood had been mixed in with


41

the wet clay. Although the clay initially seemed to become more plastic and easy to work with upon the addition of blood, after only a few seconds, the clay would immediately begin to harden. This outcome was reached regardless of which clay was used. (The experiments were undertaken with the mountain and beach clays in Tununak as well as alluvial clays collected from southern Nevada.) It also resulted regardless of the type of blood used; the experiments were repeated with (frozen and thawed) seal blood, (frozen and thawed) duck blood, and (fresh) pig blood. Based on the results of these studies, we suspect that the ethnographic reports of blood being added to wet pastes are in error. Our experiments also indicated that pottery manufacture becomes more problematic when oils are added to wet pastes, though the degree of the problems varied depending on the clays used and the drying conditions. For all of the clays, the addition of seal oil made the pastes more difficult to shape and more prone to cracking during the drying process. However, the cracking problem was especially pronounced for the highly plastic Alaskan mountain clays and in those instances when the drying process was fairly rapid. Cracking was still a problem, but less so, when oil was added to the less plastic clays or when the drying process was slowed down by a humid environment. From these findings, we conclude that the addition of oil to pastes may have been possible in some conditions, but only when low-shrinkage clays were used and when the drying process was allowed to proceed very slowly. In contrast to these findings, we experienced no problems when adding seal oil or animal blood to the surface of the already shaped pots. In fact, after firing, seal oil and seal blood produced a coating that was visually identical to the coatings seen on prehistoric sherds (Harry, K.G., Frink, L., Swink, C., and Dangerfield, C. Ms. on file, Department of Anthropology & Ethnic Studies, University of Nevada Las Vegas, Las Vegas). Additionally, we experimented with filling fired but porous pots with an oil and water mixture and allowing this liquid to permeate the vessel walls. We found that, through this process, the pores of the vessels were eventually plugged and even highly porous vessels would ultimately become capable of holding liquid (for a description of these experiments, see Harry and Frink 2009). Armed with this knowledge, we were ready to proceed with the more structured laboratory tests.

Controlled Laboratory Tests The controlled experiments were undertaken to evaluate the effects of the two surface treatments described above (that is, the application of seal oil and seal blood) on the strength and heating capabilities of pottery. In this section, we describe the methods used in these experiments and the results obtained. Effect of Surface Treatment on Ceramic Strength To evaluate the effects of surface treatment on ceramic strength, a series of test tiles were made and the force required to break them was measured. All test tiles were produced using a clay we have termed “Moapa Red,” collected from the vicinity of Logandale, Nevada. The Alaskan clays were avoided because of the cracking problems associated with their use. The pastes were made from a mixture of 80%

42

Harry et al.

dry clay with 10% fine sand and 10% coarse sand. All sands were collected from the beaches of Tununak.2 After the pastes were thoroughly mixed, they were rolled to a uniform thickness of 1 cm and cut into circular shapes measuring 4 cm in diameter. All tiles were fired in an oxidizing kiln to 650°C with a 20-min holding time at that temperature. This firing process produced tiles that were “ceramic” in the sense that they did not disintegrate when exposed to water. However, the tiles were still relatively soft and can best be characterized as low-fired. Five sets of test tiles, each containing ten specimens, were produced. The five groups of tiles included samples having (a) no surface treatment at all, (b) an application of seal oil before firing, (c) an application of seal oil after firing, (d) applications of seal oil both before and after the tiles were fired, and (e) the application of seal blood before the firing process. When seal oil was applied before the firing process, we found that it was rapidly absorbed. Therefore, when the oil was applied before the tiles were fired, three coats of oil were added. When seal oil was applied after the firing process, the oil was not as easily absorbed into the tile and, therefore, only a single coat of oil was applied. For the single set of tiles in which seal blood was applied, the blood was applied at the leather-hard stage while the clay was still slightly wet. This timing was selected because previous experiments had demonstrated that blood does not adhere well to very wet clay. Nor will it adhere to bone dry clays as it flakes off during the drying process. Despite the fact that the Arctic cooking pots would have been fired or hardened in open fires where neutral or reducing atmospheres would have prevailed, for the purposes of this experiment, we elected to fire the tiles in an oxidizing kiln. Reducing kilns or open fires were avoided because of a lack of control over their firing conditions. In previous studies, for example, we have observed that when using a “raku” or reducing kiln, at any given time during the firing process, temperatures in the chamber can vary by as much as 200°C over only a few inches of space. Although we have not measured the amount of oxygen in the kiln, we suspect that it can vary just as widely. Because our goal was to isolate the effect of using seal oil and seal blood as surface treatments, we considered it important to keep all other variables—including the paste, firing temperature, and firing atmosphere—as identical as possible for all of the tiles. Accordingly, we elected to use the oxidizing kiln where control over the firing conditions could be maintained. Unfortunately, an unavoidable drawback to this method is that most of the seal blood reacted with the atmosphere and oxidized off the tiles during the firing process. In fact, after firing, no visible blood was left on the surface of the tiles. Despite this fact, the results of the experiments suggest that either some blood remained in or on the tiles or the blood had altered the clay in some way (see discussion below). Once manufacture of the tiles was completed, their tensile strengths were measured using the ball-on-three-ball test described by Neupert (1994). These experiments were conducted using a United SSTM-1 Universal Testing Machine. A ball-on-three-ball 2

In this and the following experiments, no attempts were made to replicate Thule paste recipes. Because our goal was to isolate the effects of specific variables, our experimental design did not require that we use any particular paste recipe but only that the paste recipe be consistent within any one set of experiments. In point of fact, because of the great variability exhibited in Thule ceramic pastes, such replication would have been nearly impossible.


43

biaxial test fixture was fabricated to hold the test tiles and the experiments were conducted under quasistatic conditions. The results of the strength tests are presented in Table I and graphically illustrated in Fig. 3. These data demonstrate that the application of seal oil and seal blood to the surfaces of ceramics increases their strength. The greatest increases occurred when seal oil was applied both before and after the firing process. Surprisingly, given that most or all of the seal blood was oxidized away during the firing process, the application of the blood also resulted in increased strength. We would expect this increase in strength to be even more pronounced for pottery fired under neutral or reducing atmospheric conditions. Effect of Surface Treatment on Thermal Properties The second set of controlled laboratory experiments was designed to evaluate the effect of the two types of surface treatments on the thermal properties of vessels. To conduct this experiment, small Thule-shaped vessels were made in each of the same five categories as before. In this instance, three vessels were made for each category. The vessels were made using the same paste recipes as described for the test tiles and were again fired in an oxidizing kiln to 650°C with a 20-min soak. To ensure consistency of shape and thickness, the vessels were made by molding the clay inside of commercially purchased terra cotta flower pots. The vessels averaged 7.1 cm tall by 7.5 cm in diameter at the outside rim, which was the widest part of the vessel. The walls averaged 1.5 cm in thickness, a measurement similar to that found in Thule vessels. The capacity of the vessels ranged from 45 to 50 mL. After the replicas had been completed and fired, each pot was filled with 40 mL of distilled water and placed on one of three hotplates set to the same temperature setting. To ensure that the surfaces of the hotplates were identical, their temperatures were measured prior to placing the pots on them. Subsequently, the time taken for the contents of each pot to reach a boil was recorded. The results of this experiment are presented in Table II and Fig. 4. These data indicate that the application of seal blood and seal oil substantially decreased the time needed to bring the water to a boil. The effect was greatest with the application of seal blood, again despite the fact that most of the blood appeared to have oxidized off of the pots during firing. Producing an Unfired Cooking Vessel The results of these experiments suggest that, when applied to a surface of a vessel, seal oil and seal blood can strengthen weak vessels and improve their heating Table I The Mean and Standard Deviation of Load Required to Break Test Tiles Group No surface treatment Prefire seal oil Postfire seal oil Prefire and postfire seal oil Prefire seal blood

Sample size

Mean load (kg)

Standard deviation

10 10 10 10 10

33.6 38.5 44.6 49.2 41.5

3.0 2.2 2.2 7.0 2.7

44

Harry et al.

Fig. 3 Results of the strength tests.

55

95% Cl Load (kg)

50

45

40

35

30 No surface treatment

Pre-fire seal oil

Post-fire seal oil

Pre and post-fire seal oil

Pre-fire seal blood

capabilities. To determine whether these effects were substantial enough to enable unfired vessels to be used as cooking pots, we conducted a final set of experiments. For these experiments, we produced four sets of cooking vessels. These included those with (a) no surface treatment, (b) an application of seal oil to their surfaces, (c) an application of seal blood to their surfaces, and (d) an application of both seal oil and seal blood to their surfaces. For the latter three groups of vessels, the substances were applied to all vessel surfaces, both exterior and interior. For the vessels containing seal blood applications, the coat of blood was applied first and allowed to dry before adding the coat of oil. This procedure was followed because previous experiments had demonstrated that when the oil was applied first, the blood would not adhere to the clay surface. Two sets of pots were made in each group, so that each experiment was conducted twice. All pots were made using an alluvial clay collected from southern Nevada. The paste was comprised of 75% clay, 12.5% straw, and 12.5% sand and gravel. The pots were constructed to relatively standardized sizes measuring an average of 11 cm tall by 13 cm wide at the outside rim (the widest part of the vessel) and having wall thicknesses that averaged 1 cm. All vessels were hand-formed by Andreas Charest; the surfaces were lightly smoothed by hand but otherwise no smoothing or polishing tools were used. None of the vessels were fired. Once the pots had air-dried, they were placed on a small fire and immediately filled with 400 mL of distilled water. Every effort was made to keep the fires as

Table II The Mean and Standard Deviation of the Number of Minutes Required to Bring Water to a Boil Group No surface treatment Prefire seal oil Postfire seal oil Prefire and postfire seal oil Prefire seal blood

Sample size

Mean time (min)

Standard deviation

3 3 3 3 3

19.7 16.0 14.0 13.3 10.7

3.2 4.4 1.7 1.5 2.1

How to Make an Unfired Clay Cooking Pot Fig. 4 Results of the effect of surface treatment on thermal properties.

45

21

Minutes to boil

18

15

12

9

No surface treatment

Pre-fire seal oil

Post-fire seal oil

Pre and post-fire seal oil

Pre-fire seal blood

standardized as possible in size and amount of heat that they generated. In practice, thermometer readings indicated that the fires varied between 350°C and 430°C. The results, which are summarized in Table III, indicate that unfired vessels coated with both seal oil and seal blood can be used as cooking containers. However, it appears that both the oil and the blood are integral parts of this technology. The oil appears necessary to sufficiently raise the temperature of the water, whereas the blood appears to seal and protect the unfired clays from exposure to the water. Figure 5 shows the condition of the vessels after the firing experiments.

Discussion and Conclusion Although the esthetics of Arctic cooking pots pale in comparison to the craftsmanship shown in many other Arctic crafts, one cannot help but be impressed by the technological knowledge that lay behind their construction. The ability to take raw clays and shape them into unfired pots capable of holding and heating liquids demonstrates an intimate knowledge of the Arctic environment and of the mechanical properties of the various resources found in that world. The key to achieving this appears to have been in the use of oil and blood. Table III Results of Attempts to Boil Water in Unfired Vessels Surface treatment No surface treatment Seal oil only

Result

The interior and base of both pots completely disintegrated during the cooking process The water in these two pots came to a boil, but the interior of the vessel “melted” or sloughed off, making the water extremely muddy Seal blood only The water in these two pots remained clear, indicating that the interior wall of these vessels stayed intact and did not slough off into the liquid. However, the water in these vessels could not be brought to a boil, despite being kept on the fire for 30 min Seal blood and The water in these two pots came to a boil and the water remained clear, indicating seal oil that the interior wall stayed intact

46

Harry et al.

Fig. 5 Condition of cooking pots after attempts to boil water. Surface treatments of pots: top left, no treatment; top right, seal oil only; bottom left, seal blood only; bottom right, seal oil and seal blood.

In our experiments, the addition of blood had the effect of hardening and strengthening the clays and forming a protective seal between the pastes and the water. How and why blood acts this way is not a mystery; its adhesive properties have been well-known from prehistoric through modern times. Ethnographic accounts from Alaska indicate that animal blood was used as a glue to repair soapstone vessels (Murdoch 1892:526), to cement together limestone slabs (Mathiassen 1927, cited in Stimmell and Stromberg 1986:248), and to attach arrows to shafts (Cadzow 1920:19–21; Coxe 1804:246). More recently, elders from Tununak told us that they remember using seal blood to hold together the seams of wooden bowls during their manufacture. Historically, the use of blood as an adhesive was so valued in the Arctic that dried blood was kept on hand so that glue could be quickly made whenever needed. Rasmussen (1932:97) reports that, whenever an adhesive was needed, the dried blood was turned into glue by softening it in the mouth with a little water. The success of blood as a glue is attested to by its continued use in modern industry. Until about 1960, blood glue was the primary type of adhesive used for wood products. Significantly, the major advantage compared to other types of adhesives is that it was highly water resistant (Lambuth 2003). This water resistance, of course, would have been recognized by the Arctic potters and is undoubtedly a major reason why it works so well in waterproofing the clay containers. Seal oil similarly works to strengthen and waterproof the clay containers, and in addition, it appears to improve the thermal conductivity of the pots. We propose that the oil works in three ways to help low-fired pots bring liquids to a boil. First, by plugging the pores, the seal oil keeps the water from constantly evaporating or leaking out of these highly porous containers. Second, because oil is capable of reaching higher temperatures than water (before changing into a gas


47

and dissipating into heat), its presence inside the walls of the containers would help raise the temperature of the container walls, and thus help with the transference of heat from the exterior to the interior of the pot. Finally, as does the seal blood, the seal oil appears to help create a barrier between the clay and the liquid and thus help protect the unsintered clays from disintegrating. The waterproofing properties of the seal oil are still recognized by Native Alaskans today. One of the Tununak elders that we interviewed in 2005 told us that clay mixed with seal oil was used to caulk the kayaks. This mixture, he stated, was waterproof and would not come off in the ocean. The use of seal oil and seal blood appears to have been essential to the creation of unfired cooking pots. However, they would have also improved the durability and performance of underfired or low-fired vessels. As Schiffer (1990) has demonstrated, heating effectiveness is inversely related to a vessel’s permeability, and highly permeable pots may be incapable of bringing contents to a boil. By plugging the voids of the highly porous Thule vessels, the seal oil and seal blood would have substantially improved the heating capabilities of these pots. Additionally, by strengthening the pots, they would have helped them to last longer. The use of underfired or low-fired cooking containers, of course, was not limited to the Arctic. Such vessels have been reported for the Bushmen of South Africa (Sampson 1988:41), the Late Archaic cultures of the eastern United States (Skibo et al. 1989), the Blackfeet (Ewers 1945, Wissler 1910) and Sarcee (Sapir 1923) tribes of the northern Great Plains, the Salish tribes of northeastern Washington (Ray 1932), and the Paiute and Shoshone tribes of the Great Basin (Kelly 1976:37, 77– 79). Although, with the possible exception of the Salish, none of these groups would have used seal blood or seal oil in the manufacture of their vessels, it seems likely that similar substances might have been used to waterproof those vessels. The few ethnographic reports of how these vessels were constructed would seem to support this interpretation. For example, the Bushmen are reported to have coated their fibertempered, low-fired vessels with fat and blood before being fired, and after firing, they are said to have boiled blood in them to further seal the pot (Bleek and Lloyd 1911:343–347). The Blackfeet, who are said to have only fire-hardened their vessels, coated their clay containers with grease (Ewers 1945; Wissler 1910:265), and Great Basin tribes similarly coated their pottery with thick syrupy liquids made of boiled desert mallow plants (Steward 1933:266–267). Thus, although unfired cooking vessels would seem to be a contradiction of terms, such vessels were produced and used at various times throughout prehistory and history by hunter–gatherers living in various areas of the world. Although such vessels appear technologically simple, in fact, their production required substantial technological knowledge. One way to develop an understanding of how these vessels worked for people is to draw from multiple lines of evidence which attempt to answer several questions about the cooking vessels. In this paper, we ask how these pots, which tend to flaunt engineering standards, were constructed and how they performed. What we have found is that, in the Arctic, the makers of unfired and low-fired vessels employed several techniques, including the application of seal oil and blood, to ensure that the vessels were capable of carrying out their intended function.

48

Harry et al.

Acknowledgements This research was made possible by a National Science Foundation Office of Polar Programs grant (#0452900, Program Officer Anna Kerttula de Echave) along with logistic support from the United States Fish and Wildlife Service. We wish to thank Cory Dangerfield and Clint Swink for their work with reconstructive experiments in Tununak and Elisa George and David Yoder for their help with the laboratory experiments. We also would like to thank Raymond Kozak for fabricating the text fixture, Stacy Nelson for helping to conduct the strength experiments, and Hal Rager and Mark Slaughter for their help in finalizing the figures. Examination of archaeological sherds was made possible through the assistance of the faculty and staff of the Museum of the North, University of Alaska, Fairbanks, and we especially thank Don Odess, Angela Lynn, and Jim Whitney. We express our deepest thanks to the many Tununak elders we consulted with who provided information and other insights into the use of the clays and pottery from the region. Finally, we would like to thank Jelmer Eerkens, Jim Skibo, and three anonymous reviewers for their comments that helped to improve this article.

References Bleek, W. H. I., & Lloyd, L. C. (1911). Specimens of Bushman folklore. London: George Allen. Bogoras, W. (1904). The Chuckchee. Memoirs of the American Museum of Natural History 11. New York: Johnson Reprint. Cadzow, D. A. (1920). Native copper objects of the Copper Eskimo. New York: Museum of the American Indian, Heye Foundation. Collins, H. G. (1937). Archaeology of St. Lawrence Island, Alaska (vol. 96(1)). Washington, DC: Smithsonian Miscellaneous Collections. Coxe, W. (1804). Account of the Russian discoveries between Asia and America. London: Cadell and Davies. de Laguna, F. (1939). A pottery vessel from Kodiak Island, Alaska. American Antiquity, 4(4), 334–343. de Laguna, F. (1940). Eskimo lamps and pots. The Journal of the Royal Anthropological Institute of Great Britain and Ireland, 70(1), 53–76. de Laguna, F. (1947). The prehistory of Northern North America as seen from the Yukon. Menasha: Society for American Archaeology Memoirs, No. 3. de Laguna, F. (2000). Travels among the Dena: exploring Alaska’s Yukon Valley. Seattle: University of Washington Press. Eerkens, J. (2003). Residential mobility and pottery use in the western Great Basin. Current Anthropology, 44(5), 728–738. Eerkens, J. (2004). Privatization, small-seed intensification, and the origins of pottery in the Western Great Basin. American Antiquity, 69(4), 653–670. Eerkens, J. W., Neff, H., & Glascock, M. D. (2002). Ceramic production among small-scale and mobile hunters and gatherers: a case study from the southwest Great Basin. Journal of Anthropological Archaeology, 21(2), 200–229. Ewers, J. C. (1945). The case for Blackfoot pottery. American Anthropologist, 47(2), 289–299. Fienup-Riordan, A. (1975). Maraiuirvik Nunakauimi: history and development of pottery at Toksook Bay. Unpublished manuscript, Alaska Humanities Forum. Copy on file at the ANCSA Office of the Bureau of Indian Affairs, Anchorage, Alaska. Friday, J. (1983). ANCSA interview 83VAK23. Record on file, Bureau of Indian Affairs ANCSA Office, Anchorage, Alaska. Frink, L., & Harry, K. G. (2008). The beauty of “ugly” Eskimo cooking pots. American Antiquity, 73(1), 103–120. Gibson, A., & Woods, A. (1997). Prehistoric pottery for the archaeologist. London: Leicester University Press. Gordon, G. B. (1906). Notes on the Western Eskimo. Transactions, Department of Archaeology, University of Pennsylvania, 11, 1, Philadelphia. Harry, K. G., & Frink, L. (2009). The Alaskan clay cooking pot: why was it adopted? American Anthropologist, in press. Kelly, I. (1976). Paiute Indians II: Southern Paiute ethnography. New York: Garland. Lambuth, A. A. (2003). Protein adhesives for wood. In A. Pizze, & K. K. Mittal (Eds.), Handbook of adhesive technology (pp. 457–478). New York: Marcel Dekker. Mathiassen, T. (1927). Archaeology of the Central Eskimos (vol. 4, pp. 1–237). Report of the Fifth Thule Expedition, 1921–1924, Copenhagen.


49

McGhee, R. (1980). Technological change in the prehistoric Eskimo cultural tradition. Canadian Journal of Archaeology, 4, 39–52. Murdoch, J. (1892). Ethnological results of the Point Barrow Expedition (pp. 19–441). In 9th Annual Report of the Bureau of American Ethnology for the Years 1887–1888, Washington. Nelson, E. W. (1899). The Eskimo about Bering Strait. In J. W. Powell (Ed.), Eighteenth Annual Report of the Bureau of American Ethnology (pp. 3–518). Washington: Government Printing Office. Neupert, M. (1994). Strength testing archaeological ceramics: a new perspective. American Antiquity, 59 (4), 709–723. O’Leary, M. (1999). Early Yupiit–Cupiit regional groups with special reference to ceramic cooking pot designs. Paper presented at the 26th Annual Meeting of the Alaska Anthropological Association Meeting, Anchorage. Osgood, C. B. (1940). Ingalik material culture. YU-PA, 22. Oswalt, W. H. (1952). Pottery from Hooper Bay Village, Alaska. American Antiquity, 18(1), 18–29. Oswalt, W. H. (1955). Alaskan pottery: a classification and historical reconstruction. American Antiquity, 21(1), 32–43. Rasmussen, K. (1932). Intellectual culture of the copper Eskimos (vol. IX). Report of the 5th Thule expedition 1921–1924, Copenhagen, Denmark. Ray, V. (1932). Pottery on the Middle Columbia. American Anthropologist, 34(1), 127–133. Ray, D. J. (1975). The Eskimos of the Bering Strait, 1650–1898. Seattle: University of Washington Press. Reid, K. C. (1989). A materials science perspective on hunter–gatherer pottery. In G. Bronitsky (Ed.), Pottery technology: ideas and approaches (pp. 167–180). Boulder: Westview. Reid, K. C. (1990). Simmering down: a second look at Ralph Linton’s ‘North American cooking pots’. In J. M. Mack (Ed.), Hunter–gatherer pottery from the far west (pp. 7–18). Carson City: Nevada State Museum Nevada State Museum Anthropological Papers No. 23. Rodrίguez, C. (1995). Sites with early ceramics in the Caribbean Littoral of Colombia: a discussion of periodization and typologies. In W. K. Barnett, & J. Hoopes (Eds.), The emergence of pottery: technology and innovation in ancient societies (pp. 145–156). Washington, DC: Smithsonian Institution Press. Roosevelt, A. C. (1995). Early pottery in the Amazon: twenty years of scholarly obscurity. In W. K. Barnett, & J. W. Hoopes (Eds.), The emergence of pottery: technology and innovation in ancient societies (pp. 115–132). Washington, DC: Smithsonian Institution Press. Sampson, C. G. (1988). Stylistic boundaries among mobile hunter–foragers. Washington, DC: Smithsonian Institution Press. Sapir, L. (1923). A note on Sarcee pottery. American Anthropologist, 47(2), 289–299. Sassman, K. E. (1992). Gender and technology at the Archaic-Woodland “transition”. In C. Claassen (Ed.), Exploring gender through archaeology: selected papers from the 1991 Boone Conference (pp. 71–79). Madison: Prehistory. Sassman, K. E. (1995). The social contradictions of traditional and innovative cooking technologies in the prehistoric American Southwest. In W. K. Barnett, & J. W. Hoopes (Eds.), The emergence of pottery: technology and innovation in ancient societies (pp. 223–240). Washington, D.C.: Smithsonian Institution Press. Schiffer, M. B. (1990). The influence of surface treatment on heating effectiveness of ceramic vessels. Journal of Archaeological Science, 17, 373–381. Schiffer, M. B., & Skibo, J. M. (1987). Theory and experiment in the study of technological change. Current Anthropology, 28(5), 595–622. Skibo, J. M. (1992). Pottery function: a use-alternative perspective. New York: Plenum. Skibo, J. M., & Blinman, E. (1999). Exploring the origins of pottery on the Colorado Plateau. In J. M. Skibo, & G. M. Geinman (Eds.), Pottery and people: a dynamic interaction (pp. 171–183). Salt Lake City: The University of Utah Press. Skibo, J. M., Schiffer, M. B., & Reid, K. C. (1989). Organic-tempered pottery: an experimental study. American Antiquity, 54(1), 122–146. Spencer, R. F. (1959). The North Alaskan Eskimo, a study in ecology and society. Bureau of American Ethnology, Bulletin 171. Spray, Z. (2002). Alaska’s vanishing Arctic cuisine. Gastronomica: The Journal of Food and Culture, 2 (1), 39–40. Stefánsson, V. (1914). The Stefansson–Anderson Arctic Expedition of the American Museum: preliminary ethnological report. Anthropological Papers of the American Museum of Natural History, 14(1), 1– 395. Stefánsson, V. (1962). My life with the Eskimo. New York: MacMillan.

50

Harry et al.

Steward, J. H. (1933). Ethnography of the Owens Valley Paiute. University of California Publications in American Archaeology and Ethnology, 33(3), 233–350. Stimmell, C., & Stromberg, R. L. (1986). A reassessment of Thule Eskimo ceramic technology. In W. D. Kingery (Ed.), Ceramics and civilization, volume II: technology and style (pp. 237–250). Columbus, Ohio: American Ceramic Society. Velde, B., & Druc, I. C. (1999). Archaeological ceramic materials: origins and utilization. Berlin: Springer. Wissler, C. (1910). Material culture of the Blackfoot Indians. Anthropological Papers of the American Museum of Natural History, 5, 1–177 New York. Zhushchikhovskaya, I. S. (2005). Prehistoric pottery making of the Russian Far East. Oxford: ArchaeoPress BAR International Series 1434. Zhuskchikhovskaya, I. S., & Shubina, A. (2006). Pottery making and the culture history of Neolithic Sakhalan. In D. E. Dumond, & R. L. Bland (Eds.),Archaeology in Northeast Asia: on the pathway to Bering Strait (pp. 91–128). Eugene: University of Oregon Anthropological Papers No. 76.