STARCH ANALYSES REVEAL MULTIPLE FUNCTIONS OF QUARTZ "MANIOC" GRATER FLAKES FROM THE ORINOCO BASIN, VENEZUELA

Linda Perry

Linda Perry. Ph.D. Southern Illinois University Carbondale. Address: Department of Anthropology, Southern Illinois University, Carbondale IL 62901, USA. e-mail: LLperry@siu.edu

Summary

The general consensus existing among archaeologists that bitter manioc was the key crop in prehispanic agricultural systems in the lowlands of South America is based upon the indirect evidence provided by archaeological ceramic and lithic assemblages rather than the preserved remains of bitter manioc itself. Studies of microlithic grater teeth from the site of Pozo Azul Norte-1 in the Middle Orinoco basin were designed both to recover the starchy remains of bitter manioc and to assess the reliability of the assumption that these artifacts are a dependable archaeological indicator of bitter manioc use. The data indicate that "manioc" artifacts are not necessarily indicative of manioc use by the prehispanic populations of the Middle Orinoco, and that archaeological data are a more reliable indicator of prehistoric subsistence activities than are modern ethnographic analogues.

Resumen

El acuerdo general existente entre arqueólogos que la yuca amarga era el cultivo básico en los sistemas agrícolas prehispánicos de las tierras bajas de Sudamérica se basa en evidencias indirectas aportadas por arreglos cerámicos y líticos arqueológicos en lugar de restos preservados de la yuca misma. Se diseñaron estudios de dientes microlíticos de ralladores del sitio Pozo Azul Norte-1 en la cuenca del Orinoco Medio para recuperar restos de almidón de yuca amarga y evaluar la confiabilidad de la presunción que esos artefactos son un indicador arqueológico seguro del uso de la yuca amarga. Los datos indican que los artefactos para "yuca" nos son necesariamente indicativos del uso de la yuca por los pobladores prehispánicos del Orinoco Medio, y que los datos arqueológicos son un indicador más confiable de las actividades prehispánicas de subsistencia que los análogos etnográficos modernos.

Resumo

O acordo geral existente entre arqueólogos de que a mandioca era o cultivo básico nos sistemas agrícolas prehispánicos das terras baixas de Sudamérica baseia-se em evidências indiretas encontradas em arranjos cerâmicos e líticos arqueológicos em lugar de restos preservados da mandioca mesma. Se desenharam estudos de dentes microlíticos de raladores do lugar Pozo Azul Norte-1 na bacia do Orinoco Medio para recuperar restos de amido de mandioca e avaliar a confiabilidade da presunção que estes artefatos são um indicador arqueológico seguro do uso da mandioca. Os dados indicam que os artefatos para "mandioca" não são necessariamente indicativos do uso da mandioca pelos povoadores prehispánicos do Orinoco Medio, e que os dados arqueológicos são um indicador mais confiável das atividades prehispánicas de subsistência que os análogos etnográficos modernos.

KEYWORDS / Grater Teeth / Manioc / Middle Orinoco / Prehispanic / Starch Analysis /

Received: 04/30//2002. Modified: 09/10/2002. Accepted: 09/30/2002.

Introduction

The root crop bitter manioc (Manihot esculenta Crantz) is widely held to be the major source of carbohydrate exploited by prehispanic populations in lowland South America (Lathrap, 1970; Willey, 1971; Meggers, 1996) and, more specifically, in the Orinoco basin of Venezuela (Howard, 1943; Cruxent and Rouse, 1958; Lathrap, 1970; Willey, 1971; Vargas, 1979; Roosevelt, 1980). Artifacts typically associated with bitter manioc processing in modern ethnographic contexts include microlithic grater flakes and ceramic griddle fragments. Traditionally, archaeologists infer that the presence of analogous artifacts within an archaeological assemblage is a reliable indicator of bitter manioc use (Howard, 1943; Cruxent and Rouse, 1958; Lathrap, 1970; Willey, 1971; Vargas, 1979, 1981; Roosevelt, 1980; Meggers and Evans, 1983; Barse, 1989), and related models of subsistence behavior, population size, and settlement patterns are heavily reliant upon this assumption. These views, however, are based upon the indirect evidence provided by artifact assemblages rather than the preserved remains of bitter manioc itself. Therefore, it has been suggested that lithic grater flakes and ceramic griddles may represent other activities within the archaeological record (DeBoer, 1975; Pickersgill and Heiser, 1977; Dole, 1978; Vargas, 1981). To address the validity of the assumption that these artifact assemblages are evidence of manioc use, starch analyses were performed on lithic grater flakes from Pozo Azul Norte-1.

Study Site

Pozo Azul Norte-1, located in the Middle Orinoco Basin of Venezuela, is a small, stratified site on a terrace overlooking the Pozo Azul canal, a small, seasonally flooding tributary of the Orinoco river on the outskirts of metropolitan Puerto Ayacucho in southern Venezuela (Barse, 1989). Two 1 x 1m2 test units were excavated into a midden context in 1987, 1992 and 1995, and the ceramic artifacts that were recovered are representative of a single component occupation typical of the Barrancoid ceramic series of the Orinoco region (Barse, 1989). Uncalibrated radiocarbon dates from carbonized material recovered from level 3 of test unit B are AD 430 ±100 and AD 720 ±80 (Beta 52134, 52135). A single AMS date from carbonized material on a ceramic sherd from the same level yielded a date of AD 740 ±60 (Beta 63830). The lithic grater flakes analyzed in this study were excavated from this dated context, and the artifacts were bagged unwashed.

The assemblage of small quartz lithic flakes associated with ceramic griddle sherds has been interpreted as evidence that the inhabitants of Pozo Azul Norte-1 were, in fact, reliant upon bitter manioc for food (Barse, 1989). The quartz microliths at this location "...are interpreted as grater flakes - for bitter manioc grater boards… Certainly the manioc grater chips imply the presence and consumption of that cultigen" (Barse, 1989). To evaluate this assumption, starch residue analyses were performed on five quartz lithic grater flakes (Figure 1). Starch analyses were combined with use-wear studies to ensure that the residues were associated with used edges of the tools.

 

Figure 1. Scale drawings of quartz microliths analyzed for starch residues. The flakes are numbered sequentially (1-5) from the top. Scale bars: 1cm.

Methods

Starch analyses

Detailed methods from which the starch analyses were adapted are published elsewhere (Loy, 1994). The lithic tools were first examined under a dissecting microscope at 10x or 20x magnification using incident light and a black background. When areas in which starch residues would likely preserve, such as deep crevices, cracks, fissures, or holes in the stone, were located, a tiny drop of deionized water was placed on this region of the flake using a sterile disposable plastic pipetter, and this plastic tip or a metal probe was then used to agitate and scrape the material loose from the tool. The mixture of water and residue was siphoned off from the tool and redeposited on a glass slide in a 50/50 mixture of glycerin and water.

When as much starch as could be extracted by the water droplet method had been recovered from the artifact, the latter was placed in small vial with deionized water and placed in a sonic cleaner to shake loose any remaining starch granules. After sonication, the effluent was deposited in clean, new centrifuge tubes, and was processed in an Eppendorf centrifuge model 5415C for 5min at 1000RPM. The pellet was placed on a cleaned slide with a drop of glycerin mounting medium.

Starch granules were located on the slides using a compound light microscope equipped with a set of polarizing lenses. All granules were identified as starch due to the formation of the diagnostic extinction cross that occurs under cross-polarized light.

Secure identifications of archaeological starches involve the use of multiple morphological characteristics or multiple starch granules. If a population of starch granules all exhibit some or most of the typical characteristics of a given taxon, while some exhibit all the diagnostic characteristics, there is a great likelihood that the group of starch granules is derived from the same plant species. In the case of single granule identification, all diagnostic characteristics must be present to ensure a secure identification.

In addition to comparative material collected by this author, keys, descriptions and photographs published by other authors were consulted and cross-referenced to support taxonomic identification. Various dichotomous keys, photos, and descriptions are published both in print (Reichert, 1913; Piperno and Holst, 1998; Piperno et al., 2000) and in electronic media (Ugent, undated). Additionally, the author possesses a comparative collection of starches mounted on glass slides. Sixty species of plants are represented, with 74 specimens of manioc present. These starches were used for final comparative purposes. If the characteristics of the starch granule in question deviate from any source, published or otherwise, the identification is not secured.

Use-wear analyses

Methods for the creation of replica casts of the quartz grater flakes were modified from techniques published in detail elsewhere (Knutsson, 1988). The small quartz flakes were placed in a glass dish that was flooded with acetone to a depth that covered the surface of the flakes. Acetate sheets were lowered onto the flakes, and a dropper was used to pipet acetone from the dish onto the sheet until it melted into the surface features of the flake. At this point, the remaining acetone was quickly siphoned off so that the acetate would not dissolve completely and envelop the flakes rather than casting the surfaces only.

The resultant acetate replicas of the quartz grater flakes were placed under both a dissecting microscope and a compound microscope to observe if the peel had resulted in an accurate cast of the tool. The replicas were then sputter coated with gold/palladium and placed in a SEM for viewing. The entire surface of each replica was examined, and micrographs were taken using black and white Polaroid film.

Results

A total of 164 starch granules (Table I) was recovered from the grater flakes, and many of the morphotypes were readily identifiable according to diagnostic morphological features of various plant taxa (Reichert, 1913; Piperno and Holst, 1998). Four definitively identified genera and two tentatively identified taxa of plants were recovered, along with a large number of unidentified starches that likely represent residues derived from other starch-bearing plants.

 

A single starch granule on one flake probably belongs to Attalea, a Neotropical palm, while three starches from plants that have edible underground storage organs, Dioscorea sp. (yam), Maranta sp. (arrowroot), and Myrosma sp. (guapo), were securely identified (Figure 2a,b, d). A single starch granule probably derived from an underground storage organ of a species of ginger was recovered from one flake.

 

Figure 2. Scanning electron micrographs of acetate replicas of quartz surfaces. a: Freshly knapped quartz surface showing conchoidal fracture marks and smooth, glasslike surfaces. b: Use-wear on the surface of flake 1. Note that the conchoidal fracture marks have been obliterated and are replaced by rounded edges and nodulation. Magnification:1000x.

Starch granules (Figure 2c) derived from Zea mays (maize) occurred on each artifact that was examined (Table I), and their number (63) far exceeded those recovered from any other taxon. A large number (22) of tentatively identified maize starch granules was also extracted from the assemblage of artifacts. It is very likely that these granules do belong to the maize assemblage, however, too few diagnostic traits were present on these granules to allow for secure identification.

Finally, a total of 105 unidentified starch granules was recovered from the artifacts analyzed. It is not known how many species of plants are responsible for this population, although five definite types have thus far been categorized as distinctive enough to be probable diagnostics. Without comparative specimens from these unidentified types, it cannot be clear how many plants are represented in this assemblage.

The use-wear analyses revealed that all examined flakes exhibited heavy wear including rounded edges, smoothed, undulating surfaces, and pitting or nodules (Figure 3). No unused surfaces, as represented by a fresh break, could be observed on any portion of any flake.

 

Figure 3. Light micrographs of archaeological starch granules representative of the recovered assemblage. a: Starch granule of Dioscorea sp. recovered from flake 3. b: Starch granule of Maranta sp. recovered from flake 2. c: Starch granule cluster of Zea mays recovered from flake 3. d: Starch granule of Myrosma sp. recovered from flake 3.

Discussion

It is likely that the microliths examined in this study are grater teeth that were, at one time, embedded in a wooden board. First, modern manioc grater boards are constructed with lithics that are of similar size and shape to the artifacts in this study (Roth, 1916). Modern grater microliths tend to be about a centimeter in length and roughly "pyramidal" (triangular) in shape. Second, when the topography of the stone allowed for careful sampling, the extracts revealed a concentration of starch residue around the equator of the flake. This concentration may be due to the build up of plant residues at the interface of the stone and the wooden board. Starch would be rasped off at the tip of the tool, and would likely percolate in small quantities down into the hole in the wooden board that was prepared prior to the insertion of the stone chip (see Roth, 1916). If the board was washed between uses, starch remains would also be most likely to adhere at the base of the flake.

Further, the entire surface of each flake bears heavy wear. This pattern can be explained by the abrasion of plant materials against the tips of the flakes, and the movement of the bases of the flakes within and against the wood and resin matrix of the board during its use. The association of the starch residues with the wear also indicates that the plant remains are very likely associated with the use of the flakes. If these microliths do represent the archaeological remains of a wooden grater board, however, then the presence of the remains of plant material from several species, not including a trace of bitter manioc, must be resolved.

The presence of the remains of the tuberous plants, yams, arrowroot, guapo and ginger on the surfaces of the grater flakes is most easily explained. Both arrowroot and ginger produce very fibrous underground storage organs that must undergo processing to release the desired product, be it starch or pigment. If the modern use of the bitter manioc grater board is a historical phenomenon that represents a simplification of a prehispanic multiple use grating board, we would likely find the remains of many tuberous plants on the microliths therefrom.

The starch that may be derived from the cucurito palm is, perhaps, the most difficult data to explain. One would expect that a palm seed would be broken open using some sort of stone anvil. However, it has been noted that arepas, small pan-fried flatbreads, are made from the flour of the seeds by modern peoples in the Venezuelan state of Amazonas, and in the Upper Orinoco, starch from the mesocarp of the fruits is used in making cakes (Narváez and Stauffer, 1999). Perhaps a grater board was used to obtain starch from these fruits.

The maize remains are the most intriguing and unexpected data from this study, particularly because maize starch occurred on every artifact that was examined. Notably, different types of maize form distinctive morphological types in their starches (Reichert, 1913; Piperno et al., 2000). The starch types from these artifacts are from a maize type with a hard endosperm: e.g., a popcorn or a dent/flint corn rather than a flour corn. Popcorn kernels are usually removed from the cobs prior to use, and the starch evidence from Pozo Azul Norte-1 may indicate that grater boards were used prehistorically for this purpose.

The three categories of plant remains, underground storage organs, palm fruits, and seeds (maize), create a more diverse assemblage than would be expected given the usual association of microliths with bitter manioc in the lowland archaeological literature. The question, however, still remains as to why there were no remains of bitter manioc on these microliths. If these stone chips are the remains of a grater board, one would expect concentrations of manioc starch in quantities like those that were found with maize starch.

The least likely explanation is that manioc was processed using these artifacts, but the residues were worn off with further use, and the plants present are representative of the final use of the tool. Given the number of species present on the flakes, as high as 5 on flake 3, and the deep fissures and internal pockets on some of the flakes, particularly flake 1, this scenario seems unlikely.

The second possibility is that the artifact was washed after it was used for bitter manioc due to the toxicity of the plant material. If a careful cleaning occurred after the board came into contact with the prussic acid-bearing roots, then one might expect to find only the occasional stray granule. The assumption is that the board would not have been washed after use with non-toxic materials, and, again, several flakes have surface features that would trap plant residues such that washing would probably not remove them.

The final possibility is that bitter manioc was not processed with these tools. It may be that the modern manioc grater board is analogous to a prehispanic tool used for the processing of many plants that may not have included the bitter roots. Only further analysis of more grater flakes integrated within the analyses of associated artifacts and matrices from this and other sites will clarify this particular issue.

Conclusions

The food-processing artifacts from Pozo Azul Norte-1, like most assemblages of this type from lowland South America, were hypothesized to be manioc grater teeth, and the wear and patterns of starch deposition on the flakes indicate that they were, in fact, once part of a grater board. The plant remains recovered from these artifacts, however, reveal a more complex function than was previously hypothesized, involving the processing of several starchy roots including arrowroot, guapo, yam, and ginger, as well as seeds including maize and possibly palm. Given the amount of labor involved in constructing a grater board, its use in multiple tasks would seem logical, although it is unclear why manioc is not represented in the starch assemblage. It is clear, however, that the automatic assignment of manioc use to this particular type of artifact assemblage, microlithic grater teeth in association with ceramic griddle sherds, is no longer appropriate.

The assumption that bitter manioc use is indicated by the presence of small lithic flakes is common in the archaeological literature from greater Amazonia, and is based upon comparison of the archaeological tools to those used by modern indigenous groups in the region. However, the present data indicate that ethnographic analogy has not been an appropriate means to assess the function of these tools at Pozo Azul Norte-1. If these data are representative of what is likely to be found as investigations progress in greater Amazonia, archaeologists may find that their assessments of prehistoric diet have been overly simple, if not incorrect. Thus, many of the models of lowland South American population size and settlement pattern that have relied upon the presumption of bitter manioc use may require revision.

ACKNOWLEDGEMENTS

This work was supported by a grant from the National Science Foundation. I. Shimada, D. Pearsall, J. Hill, D. Ugent and P. Rice provided comments on the manuscript. W. Barse provided the author with a selection of lithic artifacts from Pozo Azul Norte-1.

REFERENCES

 

1. Barse WP ( 1989) A Preliminary Archaeological Sequence in the Upper Orinoco Valley, Territorio Federal Amazonas, Venezuela. PhD Dissertation. The Catholic University of America. 715 pp.        [ Links ]

2. Cruxent JM, Rouse I ( 1958) An Archaeological Chronology of Venezuela. Two Volumes. Pan American Union Social Science Monographs. Washington DC. 277 pp.        [ Links ]

3. DeBoer W ( 1975) The Archaeological Evidence for Manioc Cultivation: A Cautionary Note. American Antiquity 40: 419-433.        [ Links ]

4. Dole GE (1978) The Use of Manioc Among the Kuikuru: Some Interpretations. In Ford RI (Ed) The Nature and Status of Ethnobotany. Museum of Anthropology, University of Michigan Anthropological Papers Nº 67. pp. 217-247.        [ Links ]

5. Howard G ( 1943) Excavations at Ronquin, Venezuela. Yale University Publications in Anthropology Nº 28. Yale University Press, New Haven. 91 pp.        [ Links ]

6. Knutsson K ( 1988) Patterns of Tool Use. Scanning Electron Microscopy of Experimental Quartz Tools. Societas Archaeologica Upsaliensis. Uppsala. 114 pp.        [ Links ]

7. Lathrap DW ( 1970) The Upper Amazon. Thames and Hudson, London. 256 pp.        [ Links ]

8. Loy TH (1994) Methods in the Analysis of Starch Residues on Prehistoric Stone Tools. In Hather JG (Ed) Tropical Archaeobotany: Applications and New Developments. Routledge. London pp. 86-113.        [ Links ]

9. Meggers BJ ( 1996) Amazonia. Man and Culture in a Counterfeit Paradise. Rev. Edit. Smithsonian Institution Press, Washington. 214 pp.        [ Links ]

10. Meggers BJ, Evans C (1983) Lowland South America and the Antilles. In Jennings JD (Ed) Ancient South Americans. Freeman, New York. pp 287-335.        [ Links ]

11. Narváez A, Stauffer F ( 1999) Products Derived from Palms at the Puerto Ayacucho Markets in Amazonas State, Venezuela. Palms 43: 122-129.        [ Links ]

12. Pickersgill B, Heiser CBJr (1977) Origins and Distribution of Plants Domesticated in the New World Tropics. In Reed CA (Ed) Origins of Agriculture. Mouton Publishers, The Hague. pp 803-835.        [ Links ]

13. Piperno DR, Holst I ( 1998) The Presence of Starch Grains on Prehistoric Stone Tools from the Humid Neotropics: Indications of Early Tuber Use and Agriculture in Panama. J. Archaeol. Sci. 25: 765-776.        [ Links ]

14. Piperno DR, Ranere AJ, Holst I, Hansell P ( 2000) Starch Grains Reveal Early Root Crop Horticulture in the Panamanian Tropical Forest. Nature 407: 894-897.        [ Links ]

15. Reichert ET ( 1913) The Differentiation and Specificity of Starches in Relation to Genera, Species, Etc. Carnegie Institution, Washington DC. 2 Vol, 102. 918 pp.        [ Links ]

16. Roosevelt AC ( 1980) Parmana: Prehistoric Maize and Manioc Subsistence along the Amazon and Orinoco. Academic Press, New York. 320 pp.        [ Links ]

17. Roth E ( 1916) The Arts, Crafts, and Customs of the Guiana Indians. Bureau of American Ethnology Publication Nº 18. 745 pp.        [ Links ]

18. Ugent D ( undated) www.siu.edu/~ebl/amylose.htm.        [ Links ]

19. Vargas I ( 1979) La Tradición Saladoide del Oriente de Venezuela: La Fase Cuartel. Estudios, Monografías y Ensayos Nº 5. Biblioteca de la Academia Nacional de la Historia, Caracas. 326 pp.        [ Links ]

20. Vargas I ( 1981) Investigaciones Arqueológicas en Parmana: los Sitios de La Gruta y Ronquin, Estado Guarico, Venezuela. Estudios, Monografías y Ensayos Nº 20. Biblioteca de la Academia Nacional de la Historia, Caracas. 574 pp.        [ Links ]

21. Willey GR ( 1971) An Introduction to American Archaeology. Vol. 2: South America. Prentice Hall, New Jersey. 252 pp.        [ Links ]