DDOOmo

Smoothed Total Nuse = 336 age = Ndiscard = 40 uselife =

innrinn c

Large

Large

3 5

10 15 20 25 30 35 40 45

Extra Large

. rui n L J

Nuse = 67 age = 11.3

D 5 10 15 20 25 30 35 40 45

'jM

nnn„Íl Ílnr-I n

3 5 10 15 20 25 30 35 40 45

Smoothed Total Nuse= 157 age= 9.7

Ndiscard = 11 uselife = 14.4

Smoothed Total Nuse= 157 age= 9.7

Ndiscard = 11 uselife = 14.4

Innmnnrhn

Age(yr)

Figure 7. The number of in-use (a) ceramic and (b) metal ollas are plotted as a function of vessel age. From top to bottom the plots are for small, medium, family size, and the total olla assemblage. On each plot are indicated the number of in-use vessels, the mean age of in-use vessels, the number of discarded vessels per year (as inferred from the difference in year-to-year number in use, Equation 2), and the uselife. Included at the bottom of (b) is a smoothed plot of the total metal olla assemblage.

sels (<10 percent metal) and 50 percent or more ceramic vessels (< 50 percent metal), suggesting little impact of metal usage on ceramic vessel uselife in the Mantaro survey data (Tables 5 and 6).

The age distribution of in-use metal ollas was calculated for three size classes (small, medium, and family), comparable in volume to the ceramic ollas with the same size designations (Figure 7b). The metal olla age distribution is little changed if selected for households using 50 percent or more metal vessels, suggesting that the presence of ceramic vessels does not significantly impact the usage of metal vessels. A smoothed version of the total metal olla usage also was calculated (Figure 7b). Since metal vessel

Figure 8. The number of large (a) ceramic and (b) metal ollas in use, as a function of vessel age. From top to bottom, the plots are for large (grande), extra large (fiesta grande), total large olla assemblage, and a smoothed version of the total assemblage. The number of in-use vessels, the mean age of in-use vessels, the number of discarded vessels per year, and the uselife are indicated.

ages may be as high as 20 to 30 years, informant memory plays an important role in determining their distribution. In particular, there is a tendency to round age estimates of older vessels to the nearest 5-year mark. By smoothing the age distribution, we counteract this tendency; we apply a Gaussian filter of 10 years that preserves the total number of vessels, but disperses the peaks in the distribution. Uselife is best derived from the smoothed data using Equation 4, assuming that no discard has occurred between time periods when the number of vessels in use is constant or increases with age. For the 336 metal ollas in use, the mean age is 6.8 years (Table 5); their uselife is 8.4 years, yielding a yearly discard of 40 metal ollas (Table 6).

Table 6. Annual Discard and Uselife of Ceramic and Metal Ollas

_Ceramic_ _Metal_

Household Number Number _Olla Large Olla__Olla__Large 011a

Size Families Persons Discarded Uselife2 Discarded Uselife2 Discarded Uselife2 Discarded Uselife2

_Ceramic_ _Metal_

Household Number Number _Olla Large Olla__Olla__Large 011a

Size Families Persons Discarded Uselife2 Discarded Uselife2 Discarded Uselife2 Discarded Uselife2

2

17

34

16

2.62

.5

22

4

6.4

.3

17.0

3

22

66

19

2.52

1.1

21

4

8.0

1.3

15.0

4

24

96

26

2.08

1.1

35

6

8.0

1.2

8.3

5

38

190

48

2.37

1.9

28

7

8.4

1.9

15.0

6

29

174

34

2.32

1.1

38

5

14.0

2.1

17.0

7

16

112

20

2.15

.7

30

5

6.4

.6

13.0

8

27

216

31

2.58

1.2

28

5

8.3

2.0

11.0

9

12

108

18

2.22

.6

17

3

5.4

.8

16.0

>10b

13

137

18

2.44

.5

30

1

13.0

.8

20.0

<10% Metal

36

183

46

2.37

1.9

35

--

--

--

--

<50% Metal

151

858

193

2.38

7.9

29

24

7.0

4.1

14.0

>50% Metal

48

275

37

2.45

.6

27

16

10.0

6.8

15.0

Straw

153

882

190

2.37

7.0

29

25

7.4

5.7

15.0

Kerosene

39

212

38

2.32

1.1

28

10

10.0

4.1

14.0

Gas

7

40

2

2.67

.5

22

5

9.8

1.2

12.0

TOTAL

199

1134

230

2.37

8.6

29

40

8.4

11.0

14.0

a Years b Includes size = 11 (N = 1) and size = 12 (N = 3)

The age distributions for large ceramic ollas (grandes and fiestas grandes) encompasses a greater time span, since these vessels tend to be used infrequently and/or only for annual fiestas (Figure 8a). Their advanced ages and informant memory limitations again create peaks in the distribution at 5- and 10-year increments that are removed by smoothing the total. For the 247 large ceramic ollas, the mean age is 17 years and the uselife is 29 years, resulting in 8.6 large ollas discarded per year. That the mean age and uselife of large ceramic ollas differ significantly suggests that they may be subject to wear, rather than suffering purely accidental breakage. This uselife estimate may be too high, however, because of extremely old vessels included in the survey that were effectively removed from service. It also is possible that large metal ollas are in the process of replacing ceramics. Large metal ollas (N = 157) have a younger mean age (9.7 years) and a shorter uselife (14 years) compared to similarly sized ceramic ollas (Figure 8b).

The uselife for ceramic chatas and tostaderas (Table 7 and 8) is less than for ceramic ollas. The number of vessels recorded by the use survey for chatas (N = 211) and tostaderas (N = 197) was less than half the number of recorded ollas (N = 546). The chata uselife (2.22 years) is slightly less than that of ceramic ollas, perhaps related to their use for frying foods. Large chatas occur in far fewer numbers (N = 32) than large ollas (N = 247), and they also have a

Table 7. Mean Age of Chatas and Tostaderas

Household

Chata

Age

Large Chata

Age

Tostadera

Age

Size2

Number

(Year)

Number

(Year)

Number

(Year)

1

1

2.00

0

1

1.00

2

20

1.91

1

20.0

16

1.18

3

27

1.82

3

10.0

19

1.16

4

27

1.93

4

11.5

27

1.18

5

44

1.75

8

12.5

38

1.10

6

33

2.06

1

1.5

31

1.24

7

19

1.46

2

30.0

14

1.17

8

22

1.66

8

5.1

26

.99

9

8

1.13

2

11.5

12

1.06

>=10

10

1.42

3

2.0

13

1.08

TOTAL

211

1.77

32

10.2

197

1.13

a See Table 4 for number of households and persons.

a See Table 4 for number of households and persons.

Table 8. Annual Discard and Uselife of Chatas and Tostaderas

Household

Chata

Uselife

Large Chata

Uselife

Tostadera

Uselife

Size*

Discarded

(Year)

Discarded

(Year)

Number

(Year)

2

8

2.35

11

1.46

3

11

2.35

16

1.19

4

11

2.46

19

1.42

5

18

2.38

30

1.27

6

13

2.54

20

1.55

7

11

1.74

8

1.75

8

10

2.22

23

1.13

9

7

1.14

10

1.20

>=10

6

1.67

12

1.08

TOTAL

95

2.22

4

7.5

149

1.32

1 See Table 4 for number of households and persons.

1 See Table 4 for number of households and persons.

younger mean age (10.2 years). That the uselife of large chatas (7.5 years) is less than their mean age is an artifact of having few vessels, a mix of some with extreme age and many younger vessels. The tostaderas have significantly shorter uselives (1.32 years) than do ceramic ollas. Their mode of use, for frying and dry roasting, presents more thermal stress to the vessel (Varien and Mills 1997). Also, since households typically have only one tostadera, versus several ollas, the tostadera may experience more regular use.

Vessel uselife may vary with household size, dependent upon vessel morphology (Tables 6 and 8, Figure 9). For ollas there is little correlation between uselife and household size (r = -.20, N = 9, p = .60); for large ollas, the uselife offers no correlation with household size (r = .05, N = 9, p = .89). In contrast, the uselife of moderate-sized (small and medium) chatas is strongly correlated with household size (r = -.72, N = 9, p = .03); uselife decreases in larger households by - 12 year/person. This is a significant effect, for it suggests that chatas in small families have uselives that are one year longer than those in large families, consistent with the presence of fewer chatas in large families, as discussed earlier. For tostaderas, there also is correlation between shorter uselife and larger household size (r = -.32, N = 9, p=.39), with a slope of -.03 year/person. Large family size reduces the already short tostadera lifespan to have the lowest value for any item in the ceramic assemblage (<1.1 years).

We speculate that differences in cooking technique may explain differences in uselife (LeCount and Hagstrum 1990). In particular, an important variable may be the amount of water present during cooking, allowing for diffusion of heat and reduced thermal shock. In cases where more water is present, as for Wanka ollas used to boil tubers, the vessel use-lives tend to be longer and do not depend on household size (despite incrementally more usage in large households). In cases where little water is used, such as for chatas or tostaderas used for frying and roasting, the vessel uselives tend to be shortened, and the uselife correlates with household size. An explicit

" Olios

A

i i

"v'

V

2 4

6

8 10

i i i i i

2 4

6

8 10

~ Chotos i i

i * i

2 4

6

8 10

Tostaderos

- Nr y^Ñ/

1 1

Household Size

Figure 9. The uselife of ceramic (a) ollas, (b) large ollas, (c) chatas, and (d) tostaderas per household are plotted as a function of household size. Dashed lines give the best fit linear regression.

Household Size

Figure 9. The uselife of ceramic (a) ollas, (b) large ollas, (c) chatas, and (d) tostaderas per household are plotted as a function of household size. Dashed lines give the best fit linear regression.

statement of this relation was given by one Wanka informant who said of tostaderas, "These don't last because we don't use them with water. We replace them every year." This model may reconcile the Wanka results with those observed for the Kalinga (Tani 1994:57), where both regular-sized and large cooking pots were broken at higher rates in larger households. In the Kalinga case, vessels are used for cooking rice, meat and vegetables, foods requiring less time on the fire and less water than for boiling of tubers. The Kalinga vessels are more analogous to the Wanka use of chatas, and both show a uselife reduction in larger households.

Other important variables related to ceramic olla uselife may be the availability of metal ollas and natural gas for fuel. (For a review of ceramic versus metal vessel usage, see Arnold [1985:139,142-143].) Metal vessels clearly replace some ceramic ones and therefore may extend the age and uselife of ceramic ollas. For families with little (< 10 percent) metal, the ceramic olla mean age (1.69 years) and uselife (2.37 years) are less than that observed for families with greater metal usage (> 50 percent, mean age =1.81 years, mean uselife = 2.45 years) as well as for those families who cook with natural gas (mean age = 2.00 years, uselife = 2.67 years). Conversely, ceramic ollas subjected to a kerosene or natural gas flame may be more prone to breakage owing to thermal shock than those used over a fire. This may explain why large ceramic ollas have shorter uselives in families cooking with natural gas (Table 6). Wanka informants explicitly state that cooking over a kerosene flame is a cause for vessel breakage.

Informant-Reported Uselife

Direct informant reports on vessel discard provide a check on uselife estimated from the in-use vessels. The uselife for ceramic vessels was obtained from a survey of 223 households, reporting on the breakage of 439 vessels (370 with known ages). Information on recent breakage events was elicited from each household; however, informant memory and patience were clearly important factors in determining the accuracy of these data. Likewise, since the vessels were not available for inspection, some inconsistency in reporting vessel size and morphology may be expected. The type of fuel available for cooking (straw, kerosene, or gas) also was noted.

The cause of each pot's breakage was recorded; we classified these responses as either accidental breakage (N = 357,81 percent), service-wear breakage (N = 74, 17 percent), or unknown (N = 10, 2 percent). Wear breakage occurs during the cooking process. In many of these cases, the informant's stated cause was, "for much use," or the "bottom fell out," especially when full of food. Among accidental causes, animals are mentioned most often (N = 83)—especially those animals that have access to the kitchen such as dogs (N = 15) and cats (N = 12). Falling objects within the kitchen also are frequently mentioned (N = 69), especially wind or rain causing the kitchen roof to fall (N = 28). Children's play (N = 34) is another common cause of breakage, as is adult drinking and other behaviors (N = 33). Travel presents opportunities for breakage (N = 25), and noteworthy is the risk of transporting the vessel home after its purchase at market or fiesta (N= 11). In about one-third of the cases (N = 136), broken cooking vessels were reused for some other purpose such as roof/wall construction or storage (Deal and Hagstrum 1995). Of the two-thirds that were immediately thrown away (N = 303), locations for vessel discard included gully/ravine (N = 54), animal corral (N = 48), outside/behind house (N = 48), field (N = 36), street (N = 17), hill (N = 13), river (N = 9), kitchen/patio (N = 6), distant location of breakage (N = 4), and not specified (N = 68).

There are several types of breakage patterns that we have classified above as accidental, but could be alternatively considered use-related breakage. In particular, vessels often are broken because they are dropped while being handled (N = 30), washed (N = 23), set down or set upon too hard (N = 6), stirred too vigorously (N = 5), or other miscellaneous weak impacts (N = 49). Many of these causes might be insufficient to break or cause obvious damage to a new vessel, but they are catastrophic for an older vessel that has undergone many cycles of thermal shock (e.g., Tani 1994:60-64). Even some of the clearly accidental causes mentioned above (e.g., animals) might not have damaged new vessels. We mention this as a possible explanation for the discrepancy between the vessel age distributions, which clearly suggest wear-related breakage, and the high percentage of accidental breakage reported by informants.

Vessel uselife can be estimated from the mean age for informant-reported breakage events (Table 9). Olla uselife is reported as 2.0 years for both accidental and service-wear breakage. Variation in olla

Table 9. Informant Reported Uselife of Ollas, Chatas, and Tostaderas3

Large

Large

Olla

Uselife

Olla

Uselife

Chata

Uselife

Chata

Uselife

Tostadera

Uselife

Number

(Years)

Number

(Years)

Number

(Years)

Number

(Years)

Number

(Years)

Accident

146

2.0

25

13

61

1.7

29

5.0

38

1.7

Wear

28

2.0

3

21

10

3.1

8

2.9

22

1.9

Straw

141

1.9

25

14

59

1.7

33

5.0

52

1.7

Kerosene

30

2.4

2

21

12

3.0

4

1.2

8

2.2

Gas

3

5.3

1

1

--

--

--

--

--

--

TOTAL

174

2.0

28

14

71

1.9

37

4.6

60

1.7

a Discarded vessels with known age = 370.

a Discarded vessels with known age = 370.

uselife was observed for different fuels: 1.9 years for straw, 2.4 years for kerosene, and 5.3 years for natural gas. These numbers confirm previously discussed patterns observed when kerosene or natural gas are available as fuels. Chata uselife is reported as 1.9 years. In this case, the uselife for accidental breakage (1.7 years) is less than the uselife for service-wear breakage (3.1 years). As for ollas, chata breakage when straw is used as fuel (1.7 years) is less than when kerosene is used (3.0 years), again suggesting metal replacement. Tostaderas have an informant-reported uselife of 1.7 years for accidental breakage and 1.9 years for service-wear breakage. They have the highest percentage of reported service-wear breakage (N = 22, 37 percent). Large ollas have an overall reported uselife of 14 years, with some suggestion of increased uselife for service-wear breakage (21 years).

For all vessel types, except tostaderas, the informant-reported uselife is less than that derived from the vessel use survey. One contributing factor may be informant memory. The age of discarded vessels may have been underestimated, especially since these vessels were not present to aid in reconstructing their history (in contrast to the usage survey). Another contributing factor may be the proportion of vessels that experience breakage before or at the time they are first put into service. Wanka informants reported 11 instances of breakage during transport immediately following vessel purchase and four instances of breakage at the time of first usage. These vessels would not appear in the usage survey; therefore, use-lives derived from the usage survey may overestimate mean vessel age at the time of breakage.

The discrepancy in estimated uselife is noteworthy for large ollas, whose usage-survey uselife (29 years) and discard-survey uselife (14 years) differ by more than a factor of two. Informant memory may be invoked, as above, to suggest that the larger value is correct. However, an alternate possibility is that the usage survey included vessels that were included in the kitchen inventory, but were no longer actively being used, perhaps because of metal replacement. Even the presence of a few extremely old large ollas could skew the age distribution and therefore the estimated uselife. Based on these considerations and the overall number of discarded large ollas reported in Table 2, we see this as the best explanation for the uselife discrepancy, suggesting that the younger value (14 years) may be more accurate for large olla uselife.

Tostaderas also have a significantly different uselife derived from the usage survey than from the discard survey. In this case alone, the discard survey suggests a longer uselife (1.7 years) than predicted from the usage survey (1.32 years). We cannot explain this difference, but note that in both cases a sufficient number of vessels were reported to provide adequate statistics (N = 197 in usage survey and N = 60 in discard survey). In the analysis presented later in this paper, the usage-survey uselife (1.32 years) appears to give the most consistent result.

Was this article helpful?

0 0

Post a comment