11 The Ceramics LEE STEADMAN CERAMIC ANALYSIS AND PHASE DEFINITION quences elsewhere in the Titicaca Basin (Steadman 1995; Chavez 1992; Chaivez 1980/8 1), and in- A NALYSIS OF THE MAJORITY of ceramic artifacts volves the observation and recording of individual recovered from the Taraco Archaeological ceramic attributes (paste, color, finish, shape, etc.) Project excavations is completed, and samples are rather than the definition of a fixed set of at- now large enough to define three Chiripa phases at tributes, such as is used in a typological classifica- the site with some confidence. Currently, 5587 tion (Rowe 1959; Shepard 1956:307-318; see ceramic specimens (excluding sherds less than 1 Steadman 1995:48-50 for further discussion of cm-) have been analyzed from unmixed Chiripa differences in ceramic analysis). Individual levels, 3530 from Chiripa levels with some attribute analysis is a more sensitive and effective intrusive material, and 25,216 from the mixed, means of studying changes through time than a Tiwanaku fill, and plow zone levels. Further typological approach (Plog 1983:131-32; Plog and analysis and excavation will serve to refine the Hantman 1990:441-42) that necessarily stresses ceramic definitions and change some percentages, the similarities among ceramics rather than their but should not result in any major revisions of the differences. This type of analysis also permits the ceramic assemblages as presented here. independent study of individual ceramic attributes, The goal in undertaking this phase of the some of which may change through time more Chiripa ceramic analysis is to create a comprehen- rapidly than others, or be affected by different sive and detailed description of the ceramic social, economic, or political factors. The ceram- assemblage from each phase in the Chiripa se- ics from the mixed and plow zone levels of the quence through an attribute analysis of the ceramic Chiripa excavations were not subjected to a material. This analytical method has been used detailed attribute analysis, as this material could productively for the definition of ceramic se- not be securely used for the definition of the 52 Early Settlement at Chiripa, Bolivia 0=00= O co~~~~~~~~C 0 R~~~~~~~~ e lat~~~~~~~~~ 0 I.~ C 0 } FAtEV1 Soil Micromorphology 53 N1104 N 1+ Si: Micromorphological Sample N1100+ N1098 N 1096 l N 1094+ 1 i~~~~~~13 NIO94+ B13 9 B13 .4 ~~~~~~~~~~~~ASDID N1092+ W XJ B16 ASD17 S3: Micromorphological Sample L MTN N1090+ |L N 1088+ I B16 N1086+ S2: Micromorphological Sample N 1084 + + E1084+ +E970 IE972 974 +E976 +E978 +E980 FIGURE 18 SI -S3 thin sections from the Santiago area. 54 Early Settlement at Chiripa, Bolivia help identify organicg, phosphates, and coprolites predominance of fine sands and silts derived from (Altemuller & Van Vliet-Lanoe 1990). The carbonate and igneous rocks, most likely from the composition, form, and orientation of soil compo- Taraco Mountain range. A larger rock found in the nents are described according to internationally Monticulo mudbrick may have been imported. recognized terminology (Bullock et al. 1985; Full mineralogical analysis has not been per- Courty et al. 1989; FitzPatrick 1993). formed on thin sections from Chiripa sites. The present study appears to be the first soil micromorphological project from Andean occupa- ORGANIC REMAINS AND INORGANIC tional surfaces, and therefore comparative materi- MINERALS OF ORGANIC ORIGIN als are not available. Dr. Wendy Matthews, Fragments of bone were found in all samples however, has analyzed thin sections from occupa- from both site areas. The mudbrick sample from tional surfaces at several sites in the Near East (Ge Montfculo forms the only exception. Fine fish et al. 1993; Matthews 1992, 1995; Matthews et al. bone, often in a weathered and broken state, is 1997; Matthews & Postgate 1994), and this m material was used as a basis for comparison. The motaudn. On welpeevdsog mammal bone is visible (in thin section S2B). The lack of local reference material and small sample Santiago south profile contains the highest concen- size of this study contribute to the preliminary tration of bone and a small amount of eggshell and nature of this report. snail shell fragments. The Monticulo lenses have CONTROLS noticeably fewer fish bones. Natural profiles were exposed at two Although small charcoal fragments are found locations. The first profile was situated in shallow in most Monticulo and Santiago samples, intact soils approximately 100 m above the Llusco sector seeds are rare. Plant cells are visible in several of the site. This matrix is composed of a thin soil slides, with the highest abundance being found in that developed on gravels. Because the soils were the fill over the upper floor in Monticulo ASD 13 shallow, no thin section samples were analyzed (thin sections MH3 and MH4). Live and decayed from this context. The second profile was taken roots are well represented in the upper levels of from an untilled balk due west of the mound about both areas. Coprolite fragments are present in 200 meters, at approximately the same elevation both profiles at Santiago and at Monticulo in the * ~~~~~~~~~~~house fill from ASD 13 and the coffesponding and distance from the lake as the mound. The balk is adjacent to moder agricultural fields and is upper lenses. A whole coprolite was found in the covered by uncultivated grasses. Three thin Santiago south balk (thin section S2C). Associ- sections (called N thin sections) from this profile ated rare spherulites indicate a herbivore source were processed and analyzed as controls for for this context (Canti 1997). Spherulites were not natural processes operating in the soil. The found at Monticulo or Santiago south profile and analyzed samples are from depths of approxi- the small coprolite fragments noted there do not mately 40 cm, 60 cm, and 80 cm. The natural suggest a source. Pseudomorphic voids of plant samples reflect normal soil profile development mailer, which may be interpreted as grasses used with well-defined horizon formation and a de- as organic temper (Matthews & Postgate 1994), crease in pore size from top to bottom. A moder- were identified in ceramics and mudbricks. Dense ate level of soil fauna activity and roots are amorphous organic staining from decayed organic present. No microartifacts were identified in the material is characteristic of the fill between the control samples. Monticulo floor layers and, to a lesser degree, all of the Santiago contexts. Possible organic sources MINERALS are discussed in the summaries of each site. The soils and geomorphological processes of the Taraco Peninsula were the focus of a recent POST-DEPOSITIONAL FEATURES study by Argollo et al. (1996). In accordance with A number of features associated with post- their findings, our Monticulo and Santiago thin depositional disturbances were noted including sections show soils to be characterized by a more visible biological processes and microscopic Soil Micromorphology 55 mineralization of soils (table 2). Root, soil fauna, calcium carbonate distributed in the soil. Neofor- and other insects are the primary biological agents mation of calcite crystals on pores is clearly seen seen in thin section. Root is ubiquitous in the in the Santiago event B- 13 fabric (the melted upper levels, and relic root channels are present adobe " orange-granola"). The second process, but in low concentration in all of the slides. Soil clay alluviation, forms coatings on pores in the fauna activity is seen in the organic-rich fill upper layers at Monticulo. Third, sesquioxide beneath the Monticulo floors (see table 2) but does (iron and manganese) impregnation is well distrib- not appear to transverse the plaster floors. All uted in all slides, suggesting that these metals are Santiago deposits demonstrate a high degree of leaching into soil waters. Freeze-thaw and shrink- reworking by soil fauna. Under the microscope, swell action of clays are also likely to have the insect burrows seen in the slides appear infilled contributed to the reworking of soils identified in with microartifact-free soil translocated from thin section. above, suggesting that the burrows post-date Overall, post-depositional processes at Santiago's occupation. The combined effects of Monticulo were localized primarily to the sub- these biological processes are to substantially mix floor fills and the uppermost deposits that have the soil, encourage soil movement, and aerate soils. been more recently disturbed. Both biological and Three main types of pedogenic processes are mineralization processes were more pronounced at at work in the samples from Chiripa. First, there is Santiago than at Monticulo, with soil fauna being secondary mineral formation of amorphous the primary agent. TABLE 2 Micromorphological summary of natural post-depositional features. DISTURBANCE CHARACTERISTICS IN THIN SECTION DISTRIBUTION Root Oriented channels containing root fragments Upper samples all profiles, or if relic, infilled with moderately sorted especially Santiago; Monticulo mudbrick silt and sand Soil fauna Randomly oriented channels with rounded Upper samples all profiles and termini, dark lining on channel walls, dark through most Santiago slides; organic-rich infilling, characteristic void patterns Santiago mudbrick Burrowing Round to oval voids up to 5 cm in diameter, Lower slide Santiago south profile insects infilled by well-sorted, clean deposits (S2F) Crystalline Visible crystals (at high magnification) coating Especially noted in Santiago balk pedofeatures pores, light birefringence pattern in fine (S3A) groundmass Textural (clays) and Coatings on pores and mineral grains and small Present in different proportions in amorphous aggregates in the fine groundmass; dark all slides (sesquioxide) (sesquioxide) or yellow to red birefringent pedofeatures (clays) 56 Early Settlement at Chiripa, Bolivia although the yellow plaster floors are clearly CONSTRUCTION MATERIALS identifiable, they are too heavily disturbed to Mudbrick, plaster, and fill are the primary make observations on specific use or passive/ construction materials identified in thin sections active zones of flooring (Ge et al. 1993). A more from both sites (table 3). The intact mudbrick intensive sampling strategy, such as sampling (thin section MB) had features comparable to every 50 cm, would have permitted a closer samples from the Near East including pseudomor- assessment of the use of space. Both Floor 4 of phic voids of decayed organic matter, b-fabric ASD 14 and Floor 2 of ASD 15 appear to have (fine groundmass exhibiting a bright pattern of been water-lain in a single layer. No intact floors minerals under crossed polarized light), and mixed were sampled at Santiago, but small aggregates of but not homogenous fabric (Matthews 1992; 1995; a yellow soil fabric similar to Monticulo plasters Matthews et al. 1997). Fragments possessing were identified. This suggests that plastered similar features were found in the fill above ASD surfaces did exist but have subsequently been 13 at Monticulo. Rounded aggregates of high clay reworked and destroyed. and carbonate content were found in the lower Three fill types were analyzed at Monticulo, deposits of the Santiago south balk. These aggre- including the foundation deposit below ASD 15, gates appear to represent eroded mudbrick frag- sub-floor fills, and the "hacienda fill" above the ments in the process of decomposition. Monticulo archaeological sequence. The foundation deposits plasters have a characteristic yellow b-fabric and contain microartifacts in an organic-rich matrix. are readily identified in thin section. Subhorizon- The soil structure contains angular aggregates tal voids have been associated with trampling suggesting placement by human agency such as floors and intentional compaction of fills dumping and packing. The sub-floor fills are of a (Davidson et al. 1992; Ge et al. 1993, Matthews & relatively even thickness, exceptionally rich in Postgate 1994). Voids of this type are fbund in amorphous organics, and contain few micro- several contexts from Monticulo. Unfortunately, artifacts. This suggests a natural origin such as TABLE 3 Summary of construction materials and occupational events. CONSTRUCTION MATERIAL LOCATION CHARACTERISTICS Mudbrick ASD 13 Well-sorted with high clay content, pseudomorphic thin section S2F voids, parallel-striated b-fabric and no microartifact inclusions Plasters ASD 13 D-52 & D-83 Dense yellow birefringent fabric sandwiched between ASD 14 D-109 (and reworked by) an organic-rich fabric ASD 15 D-62 Santiago south balk thin sections Wall fill ASD 18 B-71 Charcoal, bone, coprolite, laminated aggregate, ceramic; mosaic b-fabric; reworked by soil fauna Sub-floor fill ASD 14 D-61 & D-108 Highly organic-rich, very few anthropogenic inclu- ASD 15 D-84 sions, reworked by soil fauna Foundation _______ fill___ASD __15_D-85_ inclusions, especially charcoal and fish bone Soil Micromorphology 57 cut turf. Phytolith analysis of the thin sections, good range of the floors and fills excavated in which has not been attempted at this prelimi- 1996. The foundation deposit in this sequence nary stage, may illuminate this possibility. The (ASD 15, D-85) is composed of well-mixed soils hacienda fills reflect their recent disturbance containing small fragments of anthropogenic and deposition. An additional fill was ana- materials (e.g., bone, charcoal) and appears to be lyzed from Santiago from the wall behind ASD midden. In contrast, the overlying fill (D-84) is 18 (B-75). This sample appeared in the field to exceptionally rich in decomposed organics and be a mudbrick, but its lack of characteristic contains very few anthropogenic inclusions. This mudbrick features and its high concentration of soil composition characterizes the deposits below microartifacts suggest that it is a clod of all the floors examined: ASD 14 (D-61, D-108) and midden material. ASD 15 (D-78, D-84). These dark organic deposits probably originate from decayed plant matter used in MONTiCULO the preparation of floors such as turf, reeds, or The area of Monticulo from where the matting. Phytolyth analysis would help clarify their lens series was collected (thin sections ML1-3) origin. was interpreted in the field as an enclosed, The three plaster floors examined show unroofed area outside of the structures. The greater variation than the sub-floor fills. Floor two samples analyzed are parallel to structures of ASD 15 (D-62) was an uneven yellowish plaster ASD 13 and ASD 14. The layers are remark- with very few anthropogenic inclusions. The ably undisturbed by anthropogenic activity, floors were disturbed and partially mixed with although soil fauna have disrupted the upper adjacent soils. Floor four of ASD 14 (D-6 1) was layers. They are very dense, exceptionally of a similar composition but was thinner and more well sorted, and contain a low concentration of ephemeral. Both floors appear in thin section to microartifacts (figure 19). The orientation, have been water-lain in a single layer. The upper- small size, and weathered appearance of most floor examined (ASD 13 D-52) is exception- microartifacts indicates that they were carried ally high in organics with abundant charcoal of in with the lens material. The particle size probable grassy plant origin. Coprolites and sorting and fine clay coatings on coarse sand spherulites, the products of dung, are scarce, and grains suggest deposition by water and may this may suggest the use of plants rather than dung represent fine matrix flowing down into this as fuel. The thin section of this floor revealed that area. The clean, sandy silt composition of the the burning event was situated between two fine mudbrick is similar to that of the lenses and layers of plaster (figure 20). The upper layer was erosion of mudbricks may have contributed to too fine to see in the field and under the micro- their formation. The base of this series is not scope had clay coatings on the sandy plaster laminated and the anthropogenic, mixed suggesting that it was water lain. Similar deposits composition is more representative of occupa- from Near Eastern contexts were interpreted as tional fill. part of floor maintenance (Matthews et al. In Near Eastern sites, Matthews (1992) 1997:298). This may be due to activities such as found that similar artifact-sparse contexts intentional dampening of fires, wet washing the studied under the microscope revealed ample floors, or a very thin re-plastering of the floor after evidence of human activity including imported the burning event. small aggregates and microartifacts. The The deposit above the floor (D-5 1) was rich absence of similar features with such excellent in aggregates of construction materials, charcoal, preservation in the Monticulo sequence and fragments of bone and clay. The orientation suggests that these areas were scrupulously and condition of construction materials suggest maintained, selectively used, or not used at all. that the house was burned and then collapsed. The The thin section series from the composition of microartifact inclusions appears to Monticulo house sequence is incomplete as represent midden deposits. The top of the several floor samples were damaged in trans- Monticulo house series contains a thick, poorly port. However, the samples analyzed provide a sorted layer with abundant, thick clay coats, some 58 Early Settlement at Chiripa, Bolivia of which are laminated, indicating extensive observed in the south balk profile. mechanical disturbance. These features reflect B-13 (thin section S3A) deposits also recent soil translocation consistent with fill from contain a high concentration of calcium carbonates hacienda and previous excavation activities. and show impregnation/cementation by iron and Although abandonment is difficult to clays which are likely to have contributed to the ascertain, it is interesting that no features indica- dense, broken-up nature of this deposit. The neo- tive of periodic abandonment were observed in formation of carbonate crystals and mineralization this house sequence up to the top of ASD 13. In of soils imply a strong influence of subsurface other contexts abandonment of structures is water flow. A possible source is the spring to the characterized by the formation of generally sterile southeast of the excavations. The unusual compo- layers transported by wind and water (Matthews et sition of B-13 suggests that it was imported to the al. 1997). This negative evidence may be an site as a construction material such as in indication that Monticulo deposits built up through mudbricks and/or coarse-textured flooring. No continuous use. There is also no conclusive undisturbed aggregates were identified in this evidence of microdebitage from food preparation section that would further clarify the nature of B- or craft activities in the small sample collection 13; hence, observations were made from a single analyzed. This suggests that the range of activities sample. practiced in Monticulo structures was limited or The ample evidence of earthworm activity that cleaning practices were exceptionally rigorous. from two samples of the ashy B-16 deposits (thin sections S3B & S3C) suggests that soil fauna SANTIAGO activity is largely responsible for the variable The deposits at Santiago were all highly appearance of B-16 across the site. Burning disturbed by post-depositional processes and do events are high in organics and are thus attractive not display as clear microstratigraphy as the to soil fauna; the "spongy" texture of this part of Monticulo layers. To summarize briefly, extensive B-16 appears to reflect soil fauna activities. These earthworm activity and roots contributed greatly to areas are also adjacent to intrusive burials that the disruption of Santiago deposits. Intrusive allowed soil fauna easy access to deeper deposits. burials may have encouraged these developments A more intensive sampling strategy is required to as did exposure. Burrowing creatures and inor- test this suggestion. ganic processes also contributed to the disruption. A sample collected from ASD 18 B-75 (thin Disturbance like burials and soil fauna activity section SB) was identified in the field as a possible increase the organic content of the soils and eroded mudbrick due to its aggregate nature. The probably contributed to the high amorphous sample contained abundant microartifacts and a organic staining seen in thin section. Dung and small patch of an ephemeral occupational surface. charcoal indicate other organic sources. Features specific to mudbricks, however, were not Even with such marked disturbance, micro- found. Due to the absence of these features, the scopic analysis of specific deposits did provide aggregate is probably not a mudbrick but more insight into their formation and nature. The most likely midden material deposited in clods. notable feature of the south balk profile is thin A sequence of mudbrick decomposition is section S2E which contains two distinct soil suggested from the patterns found in the thin fabrics: one dark, organic rich and the other light sections. Structural collapse is suggested by the and calcite rich. The dark soil continues as a presence of fractured mudbricks (e.g., in ASD 13 matrix throughout the other samples of the profile house fill). The intact mudbrick was in the and contains a mixture of amorphous organics and process of breaking up from the roots present. At microartifacts. In contrast, the light soil is in Santiago, coarse fragments that appear to derive rounded aggregates without microartifacts but from mudbricks have rounded edges suggesting with a concentration of salt crystals and clays, weathering and translocation. Aggregates were This composition is similar to that of mudbricks disturbed by intrusive minerals including calcium and these aggregates might be decomposing carbonate and sesquioxides. Soil fauna action mudbricks. Fine, plaster-like aggregates were also translocated the fine mudbrick fragments to the Soil Micromlorpvhology 59 . B g N | 1 . I I i . - S |~ ~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I-O _ | -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Mi ,M-. *>if of f000:g;~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~..... . FIGURE 1 9 Thin section of outside lenses on Monti'culo in plane polarized light. Note the fine bone at top. The field is 2.9 cm. long;. r~~~~~~~~~W L . _ _ __L -_s .o' .> ffi _EE E_ :_'__: '-..,k'.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.. . .. . -~~~~~~~~~~~~~~~~UA 60 EarlY Settlement at Chiripa, Bolivia point that buildings were no longer discernible. construction methods and a limited range of Santiago residents may also have aided the de- domestic practices. The initial results indicate that struction of standing structures. Matthews & more intensive sampling of the mound by taking Postgate (1994:201) found that the upper portions additional columns from structures and sampling of mudbricks exploded by the post-depositional selected features such as additional mudbricks will recrystallization of salts. Such processes may be quite productive. Sadly, Santiago does not have been active in the comparably semi-arid offer this level of resolution. However, soil Chiripa region. micromorphology has furthered the understanding of how the Santiago deposits became reworked CONCLUSION with obvious implications for the analysis of other Preliminary analysis of soil thin sections data sets. Thin sections of specific deposits at from Chiripa provided insights into the formation Santiago such as B-13 and B-16 were the most and preservation of both well-preserved and fruitful. In addition, reference samples from reworked deposits. At Monticulo, soil micromor- modem mudbrick structures should be collected phology was able to confirm ideas about the fine and analyzed to provide comparative materials lens formation in the area between structures. from local contexts. Overall, the initial results Observations on the house sequence support other from Chiripa are very promising and provide a lines of archaeological evidence that indicate good basis for future soil micromorphological uninterrupted use with maintenance of similar investigations in the region.