Identifying pottery clay from rice fields: an example from southern Luzon, the Philippines

Identifying pottery clay from rice fields: an example from southern Luzon, the Philippines
Margaret E. Becka,*, Mark A. Neupert
Journal of Archaeological Science 36 (2009) 843–849

Abstract

The general type of clay used by potters can often be identified from attributes of the finished ceramic vessel or sherd. This information is important for compositional and sourcing studies and may also shed light on the variables influencing clay choice such as social, economic, and landscape use patterns. Rice paddies are one type of clay source that are readily identifiable in archaeological ceramics. This paper describes rice paddy clays gathered during ethnoarchaeological studies of traditional potters in Para-dijon, southern Luzon, the Philippines. We analyze the effects of clay processing and vessel manufacture on these clays and find that the most diagnostic attribute of paddy soils, the iron oxide mottling, is retained in finished vessels.

Keywords: Ceramics, Compositional analysis, Paddy soils, Ethnoarchaeology

Tóm tắt nội dung

Nhận dạng đất làm gốm từ đất ruộng lúa: Một ví dụ từ miền Nam Luzon

Những người thợ làm gốm chọn đất từ những nguồn sét tùy thuộc vào nhiều yếu tố mà trong đó bao gồm độ gần về không gian, khả năng tiếp cận, thuộc tính dễ làm việc của đất sét, tính thích hợp để làm đồ gốm và kỹ nghệ làm gốm. Kiểu nguồn đất sét sử dụng, nếu không phải là nguồn hiện tại vẫn đang còn sử dụng thì chỉ có thể nhận biết qua một số đặc tính trên những sản phẩm gốm nguyên hay mảnh vỡ.

“Đất ruộng” từ những ruộng lúa nước vẫn còn được một số thợ gốm hiện nay sử dụng. Đất này bị tác động mạnh của con người và có một số đặc điểm dễ nhận biết, bao gồm những lắng đọng của ô xít sắt và kết tủa quanh gố rạ. Những đặc điểm này không thấy ở những đất không bị đọng nước trong một thời gian dài. Như sẽ thấy trong các phân tích chẩn đoán sự phân bố của ô xít sắt được lưu giữ ở một mức nào đó ngay cả trong quá trình làm đất nguyên liệu nung gốm và trong những đồ gốm thành phẩm, những điều này sẽ giúp cho các nhà khảo cổ nhận dạng việc sử dụng trong quá khứ của các nguồn đất ruộng để làm đồ gốm.

Kết quả có được về những đặc trưng đất ruộng thể hiện trong đất làm gốm trình bày trong bài này rút ra từ nghiên cứu trường hợp Paradijon (Hình 1). Đây là nơi mà một số thợ gốm vẫn sản xuất gốm theo phương pháp truyền thống. Họ khai thác đất sét từ ruộng lúa và nghiền đất bằng những chày gỗ rộng cho đến nhuyễn. Đất sét giữ những đặc trưng của đất ruộng như lắng đọng ôxít sắt.

Bài viết này trước hết mô tả đất ruộng và những tác động của con người; thợ gốm Paradijon khai thác và xử lý đất ruộng làm gốm.

Potters choose from available clay sources based on a variety of factors that include proximity and accessibility, clay working properties, desired vessel properties, and ceramic manufacturing technology (Arnold, 1985; Rice, 1987; Rye, 1981). The type of clay source used, if not the actual source, can often be identified from attributes of the finished ceramic vessel or sherd (e.g., Beck, 2001, 2006). This information can be extremely useful within more detailed compositional analyses and regional clay sampling efforts, which should be based on an understanding of the range of materials available (Beck and Neff, 2007; Neff et al., 1992). It may also be combined with other archaeological data to better under¬stand patterns of landscape use and control.

‘‘Paddy soils’’ from irrigated rice fields serve as clay sources for some potters today. They are also heavily anthropogenic in nature and have several distinctive features, including iron oxide mottling and precipitation along rice roots. These features are absent in soils that have not been flooded for extended periods. As shown in this analysis, the diagnostic iron oxide distribution is preserved to some extent even in processed pottery clay and in finished vessels, which enables archaeologists to identify the past use of paddy soils as clay sources.

The retention of diagnostic paddy soil features in pottery clay is illustrated here with a case study from Paradijon, a barangay (barrio or neighborhood) in the town of Gubat, province of Sorsogon, at the southern end of the island of Luzon in the Philippines (Fig. 1). This neighborhood is home to a number of regional pottery specialists, who still make and fire ceramic vessels using traditional, non-industrial methods (Neupert, 1999, 2000). They collect clay from rice fields and pound it with a large wooden pestle while wet to mix it evenly before use. The clays retain diagnostic attributes of paddy soils, such as iron oxide mottling, even after processing. The mixing process pulverizes some but not all of the iron oxide concentra¬tions, altering the overall color of the clay but not removing all traces of its origin.

This paper first describes paddy soils and their anthropogenic alterations, including iron mobilization and accumulation in waterlogged rice fields. It then addresses how Paradijon potters collect and process pottery clay from rice paddies before consid¬ering how processing changes clay attributes, such as color and iron oxide distribution. Analysts should be able to identify the use of these clays, once familiar with the diagnostic attributes and changes from processing.

1. Paddy soils

Fig.1. Location of Gubat within Sorsogon Province, the Philippines

Riceisnow cultivatedinmany locationsworldwide including the southeastern United States, eastern South America, west Africa, and southern Europe,but the largest rice-growingregionsbyfar areEast, South, and Southeast Asia (Moormann and van Breemen, 1978). Domesticated rice in Asia, Oryza sativa L., probably originated in China near the delta of the Yangtze Kiang, perhaps as earlyas 3300– 3400 BC (Chang, 1976:431; Morishima, 1984; Sorensen, 1986:273).

Rice may be grown without irrigation (known as ‘‘upland’’ rice cultivation), but often it is grown in purposefully flooded fields or paddies. The wild ancestor of domesticated rice, Oryza perennis, is found in swampy areas (Morishima, 1984:13), and O. sativa is also adapted to flooded conditions. This paper is focused on pottery clays from irrigated rice fields, because flooded conditions produce characteristic iron oxide mottling as well as lenses of clay from particle size sorting. These fields generally have fine loamy to fine clayey soils and are not fragmental (with over 35% of particles over 2 mm) at or near the surface, because coarser soils require more nutrients and water than finer soils and generally are less productive (Moormann and van Breemen, 1978:108–110).

Irrigated rice fields are distinctively modified by various agri¬cultural practices (Dudal, 1968; ICOMANTH, 2007), and these modifications can be used to identify them as such in the archae¬ological record, long after abandonment (Barnes, 1990). One major category of changes occurs when soil is moved to create level surfaces and hold water in fields, or even to create the fields themselves. In mountainous areas, slopes may be terraced for rice cultivation by forming a retaining wall and filling the space behind the wall with soil from upslope. A spectacular example of this is the terrace system of villagers around the municipality of Banaue, Ifuago province, northern Luzon, the Philippines (Broad and Cavanagh, 1993:25; Conklin, 1980; Eder, 1982; Moormann and van Breemen, 1978:88–89). In lowland areas, fields may be leveled and then bunded so that they will retain water. Changes to the soil profile are generally minor in this case and affect only the surface, such as the formation of a plow zone (Ap horizon) (Moormann and van Breemen, 1978:88).

Once construction is complete, rice cultivation activities bring about another set of changes. The surface structure generally becomes massive due to puddling. Human and animal traffic through the fields may create a compacted subsurface horizon termed a plow pan or plow sole (Dudal, 1968) or traffic pan (Moormann and van Breemen, 1978:91), which serves to reduce water loss (Rowell, 1981:450).

Human-induced aquic conditions are termed ‘‘anthric satura¬tion’’ (Engel and Ahrens, 1995) and are considered to be the diag¬nostic feature of paddy soils (Dudal, 1968). Whether natural or anthric, the saturated conditions cause a host of associated changes to the soil profile, including the migration and accumulation of iron and manganese, base status changes, organic matter changes, and increased weathering of soil minerals (Moormann and van Breemen, 1978:93–102).

The most significant chemical change in saturated soils is the change in iron solubility (Ponnamperuna, 1972:71), and the resulting nonhomogeneous accumulation of iron oxides is the visible feature of saturated soils most likely to be preserved in pottery clay. The migration and accumulation of iron is therefore described in some detail below.

1.1. Patterns of iron accumulation in rice fields

Iron, which is released from primary minerals during weath¬ering, is present as Fe3+ under oxidizing (aerobic) conditions. Ferric iron (Fe3+) is almost entirely insoluble and therefore essentially immobile by convection or mass flow. However, iron can move with the soil solution in three situations (Ellis et al., 1983:118-119):

1. when iron is chelated,

2. when Fe3+ (ferric iron) is reduced to Fe2+ (ferrous iron) under anaerobic conditions, or

3. when Fe is mobilized under very acid conditions.

Iron is considerably more soluble when associated with chelates, a phenomenon associated with the formation of spodic horizons (Birkeland, 1984). However, this is not an important part of iron mobility in rice fields. The most commonly discussed factor is redox potential, although pH affects the redox potential neces¬sary to reduce iron.

Under ‘‘reducing’’ conditions, an element other than oxygen is being reduced (is accepting electrons). Fe3+ is reduced to Fe2+ in anaerobic environments by microbes using Fe3+, rather than O2, as an electron acceptor (Ponnamperuna, 1972:71). Oxidation-reduction status of a soil can be measured as pE, a measure of ‘‘electron supplying intensity’’ similar to pH, which is a measure of proton-supplying intensity (Bartlett, 1981:80). It may also be measured as the redox potential, Eh,‘‘ a measure of electron availability potential’’ (Reddy and Patrick, 1983). Positive, higher Eh values indicate more oxidizing systems (Ponnamperuna, 1972:39-40). Iron will be reduced at higher Eh values when pH is low (Gotoh and Patrick, 1974).

Ferrous iron is much more soluble and can move through the soil profile in solution. When the soil or portion of the soil contains oxygen, the iron may precipitate as an iron oxide mineral. Iron coatings on ped surfaces may occur during gleization, when Fe2+ diffuses out of the anaerobic soil aggregate interior and is oxidized (Bartlett, 1981:78). Iron precipitation is also found in the oxygen¬ated rhizosphere along plant roots (Reddy and Patrick, 1983:28). Flooding-tolerant species such as rice transport oxygen from the shoot to the roots, primarily through the aerenchymna or air spaces in the root cortex, oxygenating the rhizosphere and permitting aerobic respiration in anaerobic environments (Marschner, 1995). Deposition of iron oxides along rice roots has been studied by several authors, and is of interest in part because of its implications for rice nutrition (Chen et al., 1980; Begg et al., 1994; Kirk and Bajita, 1995; Marschner, 1995).

Dissolved iron should initially precipitate as goethite ((iFeOOH), lepidocrocite (gFeOOH), or ferrihydrite (Fe5O7(OH) x 4H2O), although other iron oxides such as hematite (|iFe2O3) may be formed later from these initial precipitates. Alternating oxidizing and reducing conditions lead to the nonhomogeneous distribution of iron oxides in the soil, which appear as mottles and concretions (Allen and Hajek, 1989; Schulze, 1989; Schwertmann, 1988, 1993; Schwertmann and Taylor, 1989).

In summary, iron oxides will not be evenly distributed throughout the sediment of periodically waterlogged deposits such as irrigated rice fields, because iron is soluble in water under these conditions. They will instead be concentrated in areas with oxygen, such as along ped surfaces and along rice roots, where they will precipitate. The result is a light-colored sediment (leached of iron) interspersed with red mottles and other concentrations of precip¬itated iron.

 Fig.2. Clay pounding  Fig.4. Collection of pottery clay from the Ramos clay source                     

The appearance and distribution of iron oxides in irrigated fields is different from that in lateritic soils and gleyed soils, also frequently found in areas of rice cultivation. In gleyed soils, iron oxides are reduced under prolonged wet, anaerobic conditions, producing a gray, bluish or greenish color instead of a red color (Brady and Weil, 1996:100). In this case the reduced iron has not been leached from the deposits. Soils referred to as ‘‘lateritic’’ are generally red soils with high clay content created by extreme weathering of the parent material (Birkeland, 1984:136). These weathering may cause extreme enrichment of aluminum or iron that should be evenly distributed.

2. Pottery clay from rice paddies in southern Luzon, the Philippines

Fig.3. Overview of the Ramos clay source, within an irigated rice field temporary removed from cultivation

The potters and pottery clays described here are from the neighborhood or barangay of Paradijon in Gubat municipality, Sorsogon province in southern Luzon. Gubat (12°55'15.63"N, 124°07'28.66"E) is located near the southwestern tip of the island of Luzon along the coast of the Philippine Sea, roughly 20 km northeast of the Bulusan volcano. It is approximately 40 km northeast of Matnog, where ferries transport people across the Bernardino Strait to Samar, the closest island in the Visayas island group.

The area of Gubat is mapped as ‘‘warm cool hillyland’’ with ‘‘[h]igh andesitic hills,’’ less than 500 m above sea level in elevation and with slopes ranging from 1 to 18% (1:250,000 Land Manage¬ment map, Philippines Department of Agriculture, Bureau of Soils and Water Management). The dry season is from April to August, with the lowest monthly precipitation recorded in May. Rainfall in this area is roughly 100 mm per month or higher starting in May and peaks in November, with about 360-400 mm precipitation. The period of lowest rainfall occurs in March and April, when the area receives approximately 40-70 mm precipitation (Morris and Rumbaoa, 1985). The mean temperature is 27 °C (Office of the Municipal Planning and Development Coordinator, 1999:6).

Gubat municipality contained almost 50,000 people in the 1995 census and is divided into 42 barrios or barangays (Office of the Municipal Planning and Development Coordinator, 1999). A local author (Sarmiento, n.d.) describes the barangay of Paradijon as ‘‘a thriving urban district. [that] is the center of pottery-making in the province of Sorsogon since prewar days,’’ and Paradijon has been the site of multiple ethnoarchaeological studies (London, 1991; Longa-cre et al., 1988; Neupert, 1999, 2000). London (1991) worked with a sample of 16 Paradijon potters in 1981, documenting the production of cooking pots, charcoal-burning stoves, and flower¬pots. According to the 1995 census, less than two percent of the 275

Table 1

Particle size distribution of raw and processed unfired clay samples.

Particle size     Aurora    Ramos   SanIgnacio       Milchoro

                (raw clay) (raw clay) (raw clay) (processed clay)

Sand                26%       44         22             36

Silt                28         6         20             14

Clay                46        50         58             50

Textural class     Clay       Clay       Clay           Clay

Table 2

Munsell color determinations of raw and processed unfired clay samples.

Sample status and portion  Aurora (raw clay) Ramos (raw clay) San Ignacio (raw clay) Milchoro (processed clay)

As collected

Matrix, dry                2.5Y 7/2 (light gray) 2.5Y 7/2 (light gray) 2.5Y 8/1 (white) 2.5Y 7/4 (pale yellow)

Matrix, moist              2.5Y 6/2 (light brownish gray) 2.5Y 5/2 (grayish brown) 2.5Y 5/2 (grayish brown) 10YR 4/4 (dark yellowish brown)

Mottles, dry              5YR 5/8 (yellowish red) 7.5YR 6/8 (reddish yellow); 5YR 5/8 (yellowish red) 7.5YR 5/8 (strong brown) 7.5YR 5/6 (strong brown)

Mottles, moist            5YR 4/6 (yellowish red) 7.5YR 5/8 (strong brown); 5YR 4/6 (yellowish red) 5YR 5/8 (yellowish red) 5YR 5/4 (reddish brown)

Ground in mortar and pestle

Matrix, dry               2.5Y 7/3 (pale yellow) 10YR 6/6 (brownish yellow) 10YR 7/3 (very pale brown) 2.5Y 7/4 (pale yellow)

Matrix, moist               2.5Y 5/3 (light olive brown) 10YR 4/4 (dark yellowish brown) 10YR 5/4 (yellowish brown) 10YR 4/4 (dark yellowish brown)

households in Paradijon still produced pottery (Office of the Municipal Planning and Development Coordinator, 1999:204:Table 3.3). Neupert (2000:252) worked with all 28 active potters during ethnoarchaeological fieldwork in 1995, and Beck encountered 20 active potters during her fieldworkin1999. There wasnodemand in 1999 for ceramic cooking vessels, as local residents now used cheap aluminum pots, so flowerpots were the primary item produced.

‘‘Clay,’’ as potters and anthropologists use the term, refers in general to the plastic soil used to form vessels (Rice, 1987). This material actually contains varying amounts of clay (clay minerals and other particles smaller than 2 m or 0.002 mm), silt (0.05– 0.002 mm), and sand (2–0.05 mm) (Buol et al., 1989:81). Although in the past some potters in Paradijon used clays from surrounding hillsides, they now exclusively use paddy clays for ceramic manu¬facture (London, 1991:187–189; Neupert, 2000:253–254). Potters name their clay sources, often after a nearby landmark or the current owner of the rice field from which the clay is collected. The sources are within 2–3 km from Paradijon and are excavated by men either related to or hired by the female potters. The owner of the rice paddy receives some form of payment for the clay, but may still resist granting permission because it takes that part of the field out of production. On the other hand, some farmers occasionally offer their fields for clay collection because this excavation helps level the field and improve irrigation.

According to the men who dig the clay, there are three textures: sandy (barasan or baras), clean and sticky (himolot or hemolot), and a mixture of sandy and sticky clay (salado) (London, 1991:187–189; Neupert, 2000:253–254). Clay color may be white, red, black, or green. Clay is prepared by pounding the clay on a pounding board with a large wooden pestle (Fig. 2), and may take about three hours for one batch of pots. This mixes the clay and may be used to mix different clays together, such as the baras and hemolot clays.

3. Methods

The data set includes three samples of raw clay, 111 samples of processed (pounded) clay, and 28 samples from finished vessels. Almost all samples (n¼ 110) of processed clay and all 28 samples from finished vessels, analyzed as petrographic thin sections, come from Neupert’s (1999, 2000) analysis. These processed clay samples represent five separate clay sources, although most (84 percent) come from one of two sources (known as Mayor and Pura). They were collected from potters’ workshops after pounding and were fired before thin-sectioning. All thin sections used in this study, whether from processed clay or finished vessels, had a sample area of at least 4.5 cm2.

Beck collected the three raw clay samples from three different clay sources (known as Aurora, Ramos, and San Ignacio) in July 1999. As observed by Neupert (1999, 2000), all clays sources used by Paradijon potters at this time were located in irrigated rice fields that had temporarily been removed from cultivation (Fig. 3). The samples were collected from active clay pits with the assistance of men from Paradijon (Fig. 4), who were there collecting clay for ceramic manufacture.2 Beck also collected one processed clay sample from a potter’s work area after it was pounded. This clay originally came from the Milchoro source, located in a field near the Aurora source and may have been similar to the Aurora sample in appearance. Analysis of these three raw clay samples and one processed unfired sample included particle size determination using the hydrometer method (Gee and Bauder, 1986) and Munsell color determinations.

4. Results

4.1. Raw clay description

The collected raw pottery clay from the Aurora, Ramos, and San Ignacio sources contains at least 40 percent clay-sized particles, and would be classified as ‘‘clay’’ by the USDA (Brady and Weil, 1996:Fig. 4.8; Table 1). The processed clay sample from the Mil-choro source does as well (see Table 1), even thought it was described by potters as salado (a naturally sandy clay which did not require the addition of more sand as temper). The white himolot clay is represented here by the Aurora and San Ignacio samples, with dry matrix colors of light gray to white (Table 2). The Ramos sample also has a light gray matrix, but in the field this color is overwhelmed by iron oxide precipitation on the ped faces. The Ramos clay is therefore a ‘‘red’’ clay. Most of the clays at the Ramos source are red, according to one local potter. No examples were seen of the black or green clays also reported by potters.

Iron oxide accumulations are present in the Aurora, Ramos, and San Ignacio samples (Figs. 5–8). They appear primarily along roots and as individual concentrations, although the Ramos sample also has deposits along ped faces (Table 3; see Fig. 6). The mottling seen here is consistent with the location of these clays in redoximorphic soils, given the mechanisms for iron oxide precipitation outlined above.

The Munsell colors of the iron oxide mottles reflect the various iron oxide minerals present (Schwertmann, 1993:53–65). The color of the mottles in these samples (see Table 2) suggests the presence of both lepidocrocite and ferrihydrite. Lepidocrocite may be 7.5YR– 5YR in hue and over 6 in value, while ferrihydrite has a similar hue but is darker and usually has a value under 6. Fe2þ in solution is necessary for the formation of lepidocrocite, which appears primarily though not exclusively as concentrations within clay-rich

2 The pits from which the pottery samples came were not profiled. An additional trip to profile the pits would have required landowner permission and some equipment not available at the time of the visit, including a Munsell color chart. The fields were flooded at the time of the visit and posed the threat of the water-borne disease schistosomiasis.

Fig.5. Sample from the Aurora source, as collected. The arraws point to examples of Iron concentration

soils in an aquic moisture regime (Schwertmann, 1993:61; Fitzpa-trick et al., 1985). Lepidocrocite forms by the slow oxidation of iron, while rapid Fe2þ oxidation forms ferrihydrite. Rapid oxidation may occur, for example, if Fe2þ ‘‘moves toward a large aerated pore with a high Eh’’ (Schwertmann, 1993:63).

Some goethite may also be present. Both goethite and lep-idocrocite appeared in experimentally-produced iron oxide coat¬ings on rice roots (Chen et al., 1980). Goethite is found under similar environmental conditions, but is favored over lepidocrocite in calcareous soils. Goethite is generally 10YR–7.5YR in hue, although high concentrations in nodules appear brown. Other iron oxides, such as hematite and maghemite, seem unlikely in these samples. Hematite is generally redder in hue than the iron oxides here. Maghemite has a hue between 2.5YR and 5YR and a lower chroma and value than goethite and hematite (Schwertmann, 1993).

4.2. The effects of processing clay and vessel manufacture

Fig.6. Sample from the Ramos source, as collected, Iron oxides cover the ped faces of this sample, leaving visible only smaller white areas from which the iron has been leached. [For interpretation of the references to colour in this figure legend, the reader is referred to the web verison of this article]

Pounding clay before vessel manufacture disperses the iron oxide concentrations, changing the overall color and altering the size and shape of remaining concentrations. This was well demonstrated when preparing the three raw clay samples for particle size analysis. Grinding in a mortar and pestle for particle size analysis changed the light gray or white matrix of the Aurora, Ramos, and San Ignacio samples to a pale yellow, brownish yellow, or very pale brown (see Table 2). Grinding also visibly dispersed all (or almost all) of the iron oxide accumulations in the Aurora, Ramos, and San Ignacio samples.

Fig.7. Open clay pit at the San Ignacio source

Only isolated concretions are present in the Milchoro sample, collected from the potter after pounding. The Milchoro sample came from a field near the Aurora source and may originally have resembled the Aurora or San Ignacio samples in matrix color and iron oxide distribution. It is likely that many of the thin deposits along roots and the smaller concretions were dispersed during pounding, producing the pale yellow matrix color of this sample (see Table 2). Not all iron oxide concentrations were completely dispersed, however; pounding left some rounded concretions as large as 1–2 mm in length (see Table 3). Grinding a dried sample in a mortar and pestle is clearly more destructive to the iron oxide concentrations than the pounding by Paradijon potters.

Fig.8. Sample from the San Ignacio, as collected. The arrows point to examples of iron concentration 

Examination of 110 processed clay samples and 28 vessels in thin section shows that most, but not all, retain similar evidence of their rice-field origins even after processing by potters. Of the processed clay samples, 80 (73 percent) have identifiable iron concretions, and 48 (44 percent) have iron concretions at least 1 mm in length. Of the vessel samples, 23 (82 percent) have identifiable iron concretions, and 9 (32 percent) have iron concre¬tions at least 1 mm in length. These concretions are rounded in shape and widely dispersed throughout the sample when present (Fig. 9).

Fig.9. Petrographic thin sections of (A) processed clay from the Mayor source (sample number MANO19) and (B) a finished vessel (sample number MAN 140) showing overall distribution of iron oxide concretions. The arrows point to examples of iron concentration

The clay samples show considerable within-source variation in matrix color and the presence and size of iron concretions, and this variation seems to exceed between-source variation. Statistical comparisons were possible with the two clay sources with the most samples, Mayor (74 samples), and Pura (14 samples). There is no significant difference (p ¼.05) between these samples in the frequency of iron concretions (chi-square ¼ 1.10 ; df ¼ 1; p ¼ .295) or the frequency of concretions at least 1 mm in length (chi-square ¼ .0359; df¼1; p¼.850).

The iron oxide dispersal probably occurs during the initial clay processing and not later in vessel manufacture. There is no statis¬tically significant difference (p¼ .05) between processed clay and finished vessels in the frequency of iron concretions (chi-square¼ 1.05; df ¼ 1; p ¼.307) or the frequency of concretions at least 1 mm in length (chi-square¼1.22; df¼ 1; p¼ .270)

5. Conclusion

Iron oxide mottling and precipitation along rice roots is the most visible characteristic of clay from rice fields, and is the result of flooding the fields. Potters in Paradijon collect clay from rice fields, and process the moist clay by pounding and mixing with a wooden pestle before using it to form vessels. Pounding alters the overall clay color by dispersing most of the iron oxide concentrations and reduces the size of remaining concentrations.

Analysts should be aware of these effects when comparing archaeological samples to potential raw materials. For example, color should probably not be compared between the pastes of archaeological samples and raw clay samples from rice paddies, as in oxidation analysis (Beck, 2006; Shepard, 1939, 1953, 1985), because processing during ceramic manufacture will have signifi¬cantly altered the color. It is difficult to predict the color change in anything but general terms; there is considerable intra-source variation in iron oxide concentrations, and therefore in the matrix color and surviving iron concretions after processing. Pounding clay may also reduce some of the visible differences between clays from different sources. Despite pronounced differences in matrix color and iron oxide distribution in the original Ramos and San Ignacio raw clay samples (and different classifications by the potters as a result), the clays were similar in matrix color and overall appearance after grinding for particle size analysis. We recommend that archaeologists grind raw clay samples before making test tiles for comparison, and realize that color variation is unlikely to be helpful in distinguishing sources.

Even after Paradijon potters process the raw clay, however, it retains some evidence of its irrigated rice field origins in the form of isolated iron concretions. These are often macroscopically visible and appear in finished vessels as well as processed clay samples. Not every sample and sherd will necessarily have these concretions, but the use of clays from irrigated rice fields should be identifiable when examining multiple sherds within a collection.

Identifying the use of paddy clays is an important initial step when conducting compositional and sourcing studies. It may also shed light on the variables influencing clay choice and source accessibility, including social, economic, and landscape use patterns. For example, paddy clay attributes would be especially significant if found in pottery from the period of early rice culti¬vation (Doherty et al., 2000; Vincent, 2003) because they are characteristic of periodically waterlogged deposits. Of course, such deposits include not only irrigated rice paddies but irrigated fields for other crops, such as taro, and natural settings such as mangrove swamps, so the ceramic analysis is best considered as only one line of evidence.

The avoidance of paddy clays may also be significant. Their collection may not be possible in certain economic contexts, because it takes some land out of cultivation. In Paradijon within the lowlands of southern Luzon, where larger landowners control multiple fields, fallow fields may be more common than in areas where most land is worked by small farmers. Even when Paradijon farmers are willing to permit collection, their clays are not available to all potters; rice field ownership and the social networks of the landowners determines which potters have access to particular clay sources (Neupert, 2000). Clearly, basic identification of clay source type is one small piece of a larger anthropological puzzle, but it can be an important starting point.

Table 3

Description of iron oxide mottles.

Aurora (raw clay)

Location: along roots plus concretions

Dimensions: root deposits 0.6 mm wide; concretions as large as 4x5 mm or 4 x 7 mm

Ramos (raw clay)

Location: along roots and on ped faces; also present as concretions Dimensions: diffuse coating covering ped face;deposit along walls of 1.5 mm wide possible root void; concretions 2-3 mm in diameter

San Ignacio (raw clay)

Location: along roots plus concretions

Dimensions: root deposits 1 mm wide; some concretions measure 2 mm in diameter, 5 x 14 mm, and 5x6 mm

Milchoro (processed clay)

Location: isolated iron concretions within matrix

Dimensions: concretions commonly rounded,measuring 1-2 mm in diameter;occasional elongated concretion 0.8 mm wide

References

Allen, B.L., Hajek, B.F., 1989. Mineral occurrence in soil environments. In: Dixon, J.B.,Weed, S.B. (Eds.), Minerals in Soil Environments. Soil Science Society of America

Book Series 1. Soil Science Society of America, Madison, Wisconsin, pp. 199–278. Arnold, D.E., 1985. Ceramic Theory and Cultural Process. Cambridge University Press, Cambridge.

Barnes, G.L., 1990. Paddy soils now and then. World Archaeology 22 (1), 1–17. Bartlett, R.J., 1981. Oxidation and reduction in aerobic soils. In: Stelly, M. (Ed.),Chemistry in the Soil Environment. Special Publication No. 40. American Society of Agronomy and Soil Science Society of America, Madison, Wisconsin,pp. 77–102.

Beck, M.E., 2001. Pottery production at the Mugler site (14CY1-A), a central plains tradition house in north-central Kansas. Plains Anthropologist 46, 5–20.

Beck, M.E., 2006. Linking finished ceramics to raw materials: oxidized color groups for lowland desert clays. Kiva 72 (1), 93–118. Beck, M.E., Neff, H., 2007. Hohokam and Patayan interaction in southwestern Ari¬zona: evidence from ceramic compositional analyses. Journal of Archaeological Science 34 (2), 289–300.

Begg, C.B., Kirk, G.J.D., MacKenzie, A.F., Neue, H.-U., 1994. Root-induced iron oxidation and pH changes in the lowland rice rhizosphere. The New Phytologist 128, 469–477.

Birkeland, P.W., 1984. Soils and Geomorphology. Oxford University Press, New York. Brady, N.C., Weil, R.R., 1996. The Nature and Properties of Soils. Prentice Hall, Upper addle River, New Jersey.

Broad, R., Cavanagh, J., 1993. Plundering Paradise: the Struggle for the Environment in the Philippines. University of California Press, Berkeley.

Buol, S.W., Hole, F.D., McCracken, R.J., 1989. Soil Genesis and Classification, third ed.Iowa State University Press, Ames.

Chang, T.-T., 1976. The origin, evolution, cultivation, dissemination, and diversifi¬cation of Asian and African rices. Euphytica 25, 425–441.

Chen, C.C., Dixon, J.B., Turner, F.T., 1980. Iron coatings on rice roots: mineralogy and quantity influencing factors. Soil Science Society of America Journal 44, 635–639.

Conklin, H.S., 1980. Ethnographic Atlas of Ifugao: a Study of Environment, Culture,and Society in Northern Luzon. Yale University Press, New Haven.

Doherty, C., Beavitt, P., Kurui, E., 2000. Recent observations of rice temper in pottery from Niah and other sites in Sarawak. Bulletin of the Indo-Pacific Prehistory

Association 20 (4), 147–152.

Dudal, R., 1968. Genesis and classification of paddy soils. In: Kovda, V.A.,Lobova, E.V. (Eds.), Geography and Classification of Soils of Asia, pp. 194–197

(Translated from Russian). Published for the United States Department of Agriculture and the National Science Foundation by the Israel Program for Scientific Translation.

Eder, J.F., 1982. No water in the terraces: agricultural stagnation and social change at anaue, Ifugao. Philippine Quarterly of Culture and Society 10, 101–116.

Ellis, B.G., Knazek, B.D., Jacobs, L.W., 1983. Movement of micronutrients in soils. In:Kral, D. (Ed.), Chemical Mobility and Reactivity in Soil Systems. SSSA Special Publication No. 11. Soil Science Society of America and American Society of

Agronomy, Madison, Wisconsin, pp. 109–122.

Engel, R.J., Ahrens, R.J., 1995. Soil Taxonomy and Anthropogenic Soils. Accessed 2/24/2003 at the International Committee on Anthropogenic Soils web page.

Available from: http://clic.cses.vt.edu/icomanth/usdanrcs.html, maintained by

John Galbraith, Department of Crop and Soil Environmental Sciences, Virginia Polytechnic Institute and State University.

Fitzpatrick, R.W., Taylor, R.M., Schwertmann, U., Childs, C.W., 1985. Occurrence and properties of lepidocrocite in some soils of New Zealand, South Africa and Australia. Australian Journal of Soil Research 23, 543–567.

Gee, G.W., Bauder, J.W., 1986. Particle-size analysis. In: Klute, A. (Ed.), Methods of Soil Analysis, Part 1, second ed. American Society of Agronomy and Soil Science Society of America, Madison, Wisconsin, pp. 383–411. Agron. Monogr. 9.

Gotoh, S., Patrick Jr., W.H., 1974. Transformation of iron in a waterlogged soil as influenced by redox potential and pH. Soil Science Society of America Proceedings 38, 66–71.

ICOMANTH (International Committee on Anthropogenic Soils), 2007. ICOMANTH home page. Available from: http://clic.cses.vt.edu/icomanth/last, updated 12/ 18/07.

Kirk, G.J.D., Bajita, J.B., 1995. Root-induced iron oxidation, pH changes, and zinc solubilization in the rhizosphere of lowland rice. The New Phytologist 131, 129–137.

London, G.A., 1991. Standardization and variation in the work of craft specialists. In: Longacre, W.A. (Ed.), Ceramic Ethnoarchaeology. The University of Arizona Press, Tucson.

Longacre, W.A., Kvamme, K.L., Kobayashi, M., 1988. Southwestern pottery stan¬dardization: an ethnoarchaeological view from the Philippines. The Kiva 53 (2), 101–112.

Marschner, H., 1995. Mineral Nutrition of Higher Plants, second ed. Academic Press, New York.

Moormann, F.R., van Breemen, N., 1978. Rice: Soil, Water, Land. International Rice Research Institute, Los Ban˜os, Philippines.

Morishima, H., 1984. Wild plants and domestication. In: Tsunoda, S., Takahashi, N. (Eds.), Biology of Rice. Elsevier, New York, pp. 3–30.

Morris, R.A., Rumbaoa, F.M., 1985. Classification of Philippine Rainfall Patterns. IRRI Research Paper Series No. 109. International Rice Research Institute, Los Ban˜os, Philippines.

Neff, H., Bove, F.J., Lou, P.B.L., Piechowski, M.F., 1992. Ceramic raw materials survey in Pacific coastal Guatemala. In: Neff, H. (Ed.), Chemical Characterization of Ceramic Pastes in Archaeology. Prehistory Press, Madison, Wisconsin, pp. 59–84.

Neupert, M.A., 1999. Potters and Politics: the Organization of Ceramic Production in Paradijon, the Philippines. Unpublished Ph.D. dissertation submitted to the Department of Anthropology, University of Arizona, Tucson.

Neupert, M.A., 2000. Clays of contention: an ethnoarchaeological study of factionalism and clay composition. Journal of Archaeological Method and Theory 7 (3), 249–272.

Office of the Municipal Planning and Development Coordinator, Municipality of Gubat, Sorsogon, 1999. Socioeconomic and Physical Profile. Bound xeroxed volume available in the Gubat municipal library, Sorsogon province, the Philippines.

Ponnamperuna, F.N., 1972. The chemistry of submerged soils. In: Brady, N.C. (Ed.), Advances in Agronomy, vol. 24. Academic Press, New York, pp. 29–96.

Reddy, K.R., Patrick Jr., W.H., 1983. Effects of aeration on reactivity and mobility of soil constituents. In: Kral, D. (Ed.), Chemical Mobility and Reactivity in Soil Systems. SSSA Special Publication No. 11. Soil Science Society of America and American Society of Agronomy, Madison, Wisconsin, pp. 11–33.

Rice, P.M., 1987. Pottery Analysis: a Sourcebook. University of Chicago Press, Chicago.

Rowell, D.L., 1981. Oxidation and reduction. In: Greenland, D.J., Hayes, M.H.B. (Eds.), The Chemistry of Soil Processes. John Wiley and Sons, Chichester, pp. 401–461.

Rye, O.S., 1981. Pottery Technology: Principles and Reconstruction. Manuals on Archeology 4. Taraxacum, Washington.

Sarmiento, E.F., n.d. This is Gubat. Unpublished manuscript available in the Gubat municipal library, Sorsogon province, the Philippines.

Schulze, D.G., 1989. An introduction to soil mineralogy. In: Dixon, J.B., Weed, S.B. (Eds.), Minerals in Soil Environments. Soil Science Society of America Book Series 1. Soil Science Society of America, Madison, Wisconsin, pp. 1–34.

Schwertmann, U., 1988. Occurrence and formation of iron oxides in various pedoenvironments. In: Stucki, J.W., Goodman, B.A., Schwertmann, U. (Eds.), Iron in Soils and Clay Minerals. NATO Advanced Science Institutes Series. D. Reidel Publishing Company, Dordrecht and Boston.

Schwertmann, U., 1993. Relations between iron oxides, soil color, and soil forma¬tion. In: Bigham, J.M., Ciolkosz, E.J. (Eds.), Soil Color. SSSA Special Publication No. 31. Soil Science Society of America, Madison, Wisconsin.

Schwertmann, U., Taylor, R.M., 1989. Iron oxides. In: Dixon, J.B., Weed, S.B. (Eds.), Minerals in Soil Environments. Soil Science Society of America Book Series No. 1, second ed. Soil Science Society of America, Madison, Wisconsin, pp. 379–438.

Shepard, A.O., 1939. Technology of La Plata pottery. In: Morris, E.H. (Ed.), Archae¬ological Studies in the La Plata District. Carnegie Institution of Washington Publication 519. Carnegie Institution, Washington, D.C., pp. 249–287.

Shepard, A.O., 1953. Notes on color and paste composition. In: Wendorf, F. (Ed.), Archaeological Studies in the Petrified Forest National Monument. Museum of Northern Arizona Bulletin 27. Museum of Northern Arizona, Flagstaff, Arizona, pp. 177–193.

Shepard, A.O., 1985. Ceramics for the Archaeologist. Braun-Brumfield, Inc., Ann Arbor, Michigan. Original printing, Carnegie Institution of Washington Publi¬cation 609, Washington, D.C.

Sorensen, P., 1986. On the problem of early rice in Southeast Asia. In: Norlund, I., Cederroth, S., Gerdin, I. (Eds.), Rice Societies: Asian Problems and Prospects. Studies on Asian Topics No. 10. Scandinavian Institute of Asian Studies. Curzon Press, the Riverdale Company, pp. 267–279.

Vincent, B., 2003. Rice in pottery: new evidence for early rice cultivation in Thailand. Bulletin of the Indo-Pacific Prehistory Association 23 (1), 51–58.