Rock Art Research & Art History Home Reinaldo Morales Jr.
Ph.D. Dissertation


6 July 2002



APPENDIX (excerpt)


The archaeological reconnaissance in the Serra da Capivara region has been concentrated around the rock art, so the extent of prehistoric occupation near the rivers and on the plains is not as fully understood. The enormous body of evidence found in the canyons along the serras nonetheless attests to a long and varied history of occupation. Paleoindian groups may have arrived as early as 48,000 years ago, as finding at Toca do Boqueirão do Sítio da Pedra Furada would appear to indicate.1 Around 12,000 years ago, at the end of the Last Glacial Maximum, there was an increase in population2 and a slow but drastic change in the environment. Hunter-gatherers exploited the region's resources until the appearance of sedentary and ceramic producing cultures, around 3000 years ago. These groups occupied the region until the Eighteenth-century when intense European colonization began.3

This summary introduces the linguistic evidence of the origins and distribution of the various Brazilian Indian groups discussed in this dissertation. Following this is a review of the archaeological evidence, the history of environmental changes, and Guidon and Pessis' proposed chronology of the rock art in the Serra da Capivara region.


Archaeology, Environment, and
the Rock Art Chronology

The Paleolithic Period

The earliest evidence of rock art comes from the extensive excavations at Pedra Furada. Forty-six small fragments of painted sandstone had spalled from the rockshelter wall, and were recovered in levels of datable material (fig. 194).4 The earliest dated occupation level, designated Pedra Furada Phase I, included a hearth with charcoal that yielded two dates, roughly 30,210 and 29,750 B.C.5 Guidon considered that a fragment of painted rock from this level was "perhaps evidence of the ancient practice of rupestral art."6 Other painted spalls were found in the context of Pedra Furada Phases II, III and IV. These were associated with dates of 27,910, 24,350, and 15,050 B.C.7 Some doubts have been raised regarding the agency of these artifacts' deposition, but they remain the earliest possible evidence of art in the Americas.8

The Early Archaic Period and Serra da Capivara Style Rock Art

The earliest evidence of Nordeste Tradition rock art is from the site of Baixão do Perna I. The site was shown to researchers in 1973, during the early years of the Franco-Brazilian Archaeological Mission in Piauí. Subsequent excavations in the 1980s revealed a buried panel of paintings (fig. 13, top). The excavation directed by Lydia Gambèri provided the oldest dates. This information was first published by Guidon in 1989, and was recently included in an essay by Pessis.9 Levels III & IV of the excavation were actually covering the panel and dated to between 3410 and 2970 B.C.10 Level V included a piece of charcoal embedded in the rock face at the base of the paintings which yielded a date of 7700 B.C.11 Lower in this level, 20 cm below the bottom of the painted panel and 40 cm above bedrock, charcoal from a hearth was dated to 8580 B.C.12 This is not an absolute date for the paintings,13 but it is strong evidence that the Nordeste Tradition is at least 9,000 years old in the area.

Sedimentary evidence from sites in Serra da Capivara suggests that beginning around 12,000 years ago there was a significant drying of the environment, corresponding to similar conditions elsewhere in Brazil after the onset of interglacial climates.14 The earliest hearth uncovered at Baixão do Perna I was considered supporting evidence of this. Periodic flooding of the rockshelter, which would have washed out the hearth, ceased occurring by 8580 B.C.15 This period saw the retreat of tropical vegetation to well watered canyons and river valleys, and the onset of savanna vegetation in the plains. Megafauna disappear from the archaeological record with the loss of these plant resources.16 Artifacts attest to a change from locally acquired and only slightly worked quartz and quartzite cobble tools to more refined and sometimes imported flint and siltite tools.17

Based primarily upon the lithics and hearth structures from several sites (but primarily Pedra Furada), Guidon identified the Serra da Capivara (10,200 - 5000 B.C.) and Serra Talhada (6050 - 4160 B.C.) Phases of material culture.18 In subsequent publications, this entire horizon would be called the Serra Talhada Phase (10,200 - 4160 B.C.).19 The lithic technology included initially unifacial local quartz and quartzite pebble tools with only rare examples of flint (similar to the Itaparica Tradition identified by Pedro Ignacio Schmitz in Central Brazil).20 Hearths were circular, outlined with pebbles.

During the latter part of this phase (6050 - 4160 B.C.), imported flint became the most widely used material for lithics, which exhibited a greater variety of tool types (including projectile points, which, in other parts of Brazil might indicate a new dart type or method of hunting).21 Hearths varied, from small circular areas surrounded by fallen blocks, to more formally organized triangular units made up of three flat rocks fixed in the ground. Eventually, the hearths became elliptical in outline, and contained fragments of bone, wood, leaves, and fruit pits. In addition to the diversity of lithic types, there were also large quantities of ochre and fragments of painted rock dating between 6000 and 4000 B.C. Rockshelters that had at one time hosted occasional hunter-gathering parties became more frequently occupied.22 Higher temperatures and more humid seasonal patterns reached a maximum between 4500 - 2000 B.C. with the onset of the Mid-Holocene Climatic Optimum.23

The archaeological evidence of Serra da Capivara Style rock art being at least 9,000 years old is convincing (Baixão da Perna I). The fragments of earlier paintings from Pedra Furada are not substantial enough to propose a stylistic affiliation, so it remains uncertain precisely how early the Nordeste Tradition appeared in Serra da Capivara.24 The Serra da Capivara Style evolved, according to Pessis, into the Serra Talhada Complex over the next few millennia, "in a process of graphic transformation which [was] not linear."25 There is no direct evidence to support the chronology of the Serra Talhada Complex rock art, only the proposal that by reading the thematic choices of the artists, their social evolution (or ethnicity) could be divided into "spatio-temporal units;" hence, associated with a archaeological phases.26

Serra Branca Style and Salitre Subtradition Rock Art

The environmental conditions of the ensuing Climatic Optimum (4500 - 2000 B.C.) certainly forced changes in adaptive strategies; hence, changes in the material culture. As Schmitz characterized it:

These environmental changes created abundant new resources for human exploitation along the rugged coast and perhaps in parts of the interior but also made traditional and densely populated locations less productive or even inhospitable. Many rock shelters intensively occupied during earlier periods in southern Piauí, Pernambuco, Minas Gerais, and Goiás were abandoned, perhaps because they became too wet, too hot, or too poorly ventilated or perhaps because they were made obsolete by the construction of perishable shelters that permitted greater mobility. It is equally possible, however, that the impoverishment of local subsistence resources forced the occupants to move elsewhere.27

These changes were in all likelihood reflected in the art, but precisely how is impossible to determine. The terminal date of the Nordeste Tradition rock art is, nonetheless, considered to be parallel with the termination of its namesake material culture around 4000 B.C.

Guidon and Pessis placed the Serra Branca Style between 5,000 and 4,000 B.C., representing the end of the Nordeste Tradition of rock painting, and an abandonment of the area by the Nordeste people.28 Pessis has since suggested various terminal dates. She wrote in 1999 that,

the paintings of the Northeast Tradition were produced over 6000 years..., [from] 12,000 [to] 6000 BP.... Beginning around 6000/5000 years BP, the dominance of the Northeast tradition ended.29

More recently Pessis proposed the following chronology:

[The] populations [that] settled in the region between 12 thousand and 3,500 years BP... are known as the Nordeste Tradition peoples. For 9 thousand years a material culture developed with techniques each time more refined.30

This is an apparent modification of Guidon's proposal that the change in material culture at the end of the Serra Talhada Phase (c. 4000 B.C.) represented a change in the ethnic groups responsible for the art (from Nordeste to Agreste peoples).31 According to Pessis here, the Serra Talhada "stylistic complex of transition" began around 7000 B.C. and, "around 6 thousand years BP [4000 B.C.] the traces of continuity of the rupestral practice [rock art?] of the Nordeste Tradition vanished, [with] those of the Agreste Tradition persisting."32

These hypotheses have yet to be corroborated by archaeological evidence (such as was available at Baixão da Perna I for the earliest dates), or by direct sampling of the paint—absolute dates. A series of absolute dates for Serra Branca Style paintings would help to confirm the hypothesis that this style was isolated to the period of the Serra Talhada material culture (no later than 4000 B.C.). Guidon originally (1986) considered the formally similar Salitre Subtradition to tentatively date between 3000 and 2000 B.C., based upon the oldest occupation evidence from the site of Morcego (2340 - 890 B.C.).33 Later publications do not address this in detail, and the Salitre Subtradition is generally relegated to the end of the Nordeste Tradition (again, c. 4000 B.C.). Laurence Ogel-Ros proposed that the Salitre Subtradition was not an isolated subtradition, but actually a final developmental stage of the Serra Branca Style.34 However, even he did not propose a chronological placement for the paintings.

The Late Archaic Period and Agreste Tradition Rock Art

A change in material culture occurred around 4000 B.C., coinciding with the culmination of the long period of climatic drying. A predominance of more roughly worked lithic technology appeared.35 The Portuguese term agreste (rough, rural, or rustic), was used by Guidon to describe a less sophisticated lithic technology compared to that of the Serra Talhada Phase. Tools were almost exclusively manufactured from locally available quartz and quartzite, lacked extensive retouching, and rarely included bifacial work (much like the early Serra da Capivara Phase identified at Pedra Furada). Guidon hypothesized that the pronounced changes in the technology indicated a new ethnic group arriving in the Serra da Capivara area around 4000 B.C. For the next 3000 years, this Agreste Tradition of material culture characterized the local populations.

The dating of the Agreste Tradition of rock art is an example of the revision of "chrono-styles" that were initially based upon correlation with material phases. The Agreste Tradition rock art is formally distinct from the Nordeste Tradition. Figures are bulbous, and more roughly delineated, frequently with very thin pigment (almost exclusively red-monochrome executed with, apparently, fingers, as opposed to the fine brush work that characterizes the Nordeste Tradition). The paintings of the Agreste Tradition were originally considered to coincide with the appearance of the Agreste phase of material culture around 4000 B.C. Drops of paint from an Agreste Tradition composition at Boa Vista I were found in the same layer as carbon that dated to 3140 B.C. This lead Guidon to conclude that the Agreste Tradition rock art dated from 4000 to 2000 B.C. in the region.36

Gamberi's 1989 excavation results from Baixão da Perna I forced a complete reevaluation of this chronology. The panel of paintings uncovered in the archaeological strata included, to some surprise, a few red monochrome figures that were formally consistent with Agreste Tradition paintings identified elsewhere by Guidon (fig. 13, bottom).37 In fact, they were near the lower limits of the painted surface—contemporary with the 9,000 year old Nordeste Tradition (Serra da Capivara Style) paintings. A chronological revision of Agreste Tradition rock art was published with the results of the excavations. While its antiquity was pushed back to 8000 B.C., its proposed terminal date remained unchallenged—around the second or first millennium B.C., in accordance with its namesake material phase.38 In this case, the associative dating of the rock art (within a defined and datable stratigraphic context) overturned a chronology based upon the isolation of a particular painting style to a particular horizon of material (primarily lithic) complexity.39 Phil Geib has cautioned against reading a change in lithic complexity as an indicator of wholesale change in the ethnic or culture groups responsible for their manufacture.40 Guidon and Pessis' general chronological order of the styles and traditions deserves further consideration (Chart 5; see also Chart 6 for the list of dates between 11,000 B.C. and A.D. 1000, and Chart 7 for the array of those dates through time).

The Development of Ceramics and Agriculture

Available climatological evidence suggests that there was a return of humidity in the period between 3000 and 1000 B.C. The vegetation "intensified extraordinarily," and was characterized by "large forests in the valleys and cerrado on hills and slopes, ...with a dry season of around 5 months and higher precipitation."41 Conditions between 1000 B.C. and A.D. 1000 continued to become more conducive to the spread of dense cerrado and semideciduous forests. The dry season became shorter and the influence of fires on the vegetation diminished.42 Material culture went through a radical change with the appearance of ceramics, around 1370 B.C. These technically refined ceramics were characterized by Guidon as having "thin walls, well fired clay, [and] varied decorations (plastic and painted)."43 Lithics recovered from these cultures showed they "used exclusively quartz and quartzite for making tools, although they used jade for making exquisite ornaments."44 By 140 B.C., peanuts, beans, and gourds were being cultivated, marking the earliest evidence of agriculture in Serra da Capivara.45

Current Caatinga Environment:
Data from the Center for Plant Diversity,
National Museum of Natural History, Smithsonian Institution

The following information is excerpted from the Center for Plant Diversity Data Sheet, "Interior Dry and Mesic Forests: CPD Site SA19 (Caatinga of Northeast Brazil):"

[See: Lleras, "Interior Dry and Mesic Forests," ]


1. Guidon and Arnaud, "Chronology of the New World." Hereafter the site is referred to as simply Pedra Furada. The author is not in the position to contest the antiquity of the archaeological evidence from Pedra Furada. The salient material is cited regarding the controversy over Guidon's work. Both the pros and cons are well published. The growing body of material evidence, and the increasing acceptance among scholars of pre-Clovis (c. 11,000 B.C.) occupations in South America nonetheless deserves mention and serious consideration.

2. Or an initial peopling.

3. Guidon and Arnaud, "Chronology of the New World," and Guidon, "Ocupações pré-históricas."

4. Pessis, "Chronology and Evolution," 44.

5. 32,160 ±1000 B.P. (GIF-6653) and 31,700 ±830 B.P. (GIF-6652). Guidon, FUMDHAM datatbase, 1998, and Guidon and Delibrias, "Carbon-14 dates," 769-771.

6. Guidon and Delibrias, "Carbon-14 dates," 769.

7. 29,860 ±650 B.P. (GIF-6651), 26,300 ±800 B.P. (GIF-6309), and 17,000 ±400 B.P. (GIF-5397). FUMDHAM database, and Guidon and Delibrias, "Carbon-14 dates," 769-771.

8. For the controversy surrounding Pedra Furada, see p. 22 of the Introduction.

9. Guidon, "Notas sobre dois sítios;" Pessis, "Chronology and Evolution."

10. 5360 ±80 B.P. (GIF 7740) and 4920 ±70 B.P. (GIF-7739). Guidon, personal communication 7 June 1998; Martin Pré-História, 75, 100; and mentioned by Prous "Dating Rock Art," 30.

11. 9650 ±100 B.P. (BETA-32972).

12. 10,530 ±110 B.P. (BETA-32971).

13. Absolute dates for rock paintings are obtained,

by coupling high vacuum techniques with a low-temperature oxygen plasma, [to] selectively remove organic carbon in rock paints without contamination from the rock substrate, accretionary coatings or atmospheric carbon dioxide. The extracted sample is then sent to an AMS [Accelerated Mass Spectrometer] for 14C dating (Chaffee, Hyman, and Rowe, "AMS 14C Dating," 67).

This provides the age of the organic material used in the paint recipe. This technique has been used to date paintings from several sites in North America, but has not produced any reliable results from samples taken in Serra da Capivara, thus far. Elsewhere the reported results should be considered with caution. Work is underway to collect samples in Serra da Capivara for absolute dating (Marvin Rowe, personal communications 1999, 2000, 2001).

14. Paleoclimate studies for the specific areas discussed in this dissertation are incomplete and, "still extremely hypothetical" (Schmitz, "Prehistoric Brazilian," 59). The information cited is a synthesis developed from Guidon's archaeological reconnaissance, and several sources addressing central and east Brazil from the Last Glacial Maximum (c. 19,000 B.C.), through the present. See: Ab'Sáber, "Domínio morfoclimático," and "Problemática da desertificação;" Adams, "South America;" Behling, "High Resolution Holocene;" Colinvaux and Oliveira, "Paleoecology and climate;" Kipnis, "Early hunter-gatherers;" and, Ledru et. al., "Last 50,000 Years."

15. Guidon and Arnaud, "Chronology of the New World," 173; and, Adams, "South America."

16. Guérin et. al., "Fauna pleistocênica," 89.

17. Guidon and Delibrias, "Carbon 14 dates," 769-771.

18. Sítio do Meio, Paraguaio, Pedra Furada, Caldeirão do Rodrigues I, Boa Vista I, Baixão do Perna I, Entrada do Pajaú, and Bojo I. Guidon, "Seqüência cultural," 139-141.

19. Guidon and Arnaud, "Chronology of the New World," and Parenti, "Questions."

20. Schmitz, "Prehistoric Brazilian," 64-71.

21. Ibid., 73.

22. Guidon, "Seqüência cultural," 138-141.

23. Guidon, "Seqüência cultural," 138-141; Behling, "High Resolution Holocene," 265-266; and, Ledru et. al., "Last 50,000 Years," 239-240.

24. Ongoing research by Dr. Guidon, FUMDHAM, and the Marvin Rowe Research Group promises to address this in the near future.

25. Pessis, "Chronology and Evolution," 45.

26. Pessis, "Identidade e Classificação," 47-53.

27. Schmitz, "Prehistoric Brazilian," 74.

28. Guidon (1998), "Ocupações pré-históricas," 48. Initially (1986), she attributed the Serra Branca Style to between 5000 and 3000 B.C. ("Seqüencia cultural," 142).

29. Pessis, "Chronology and evolution," 47. It is unclear which millenium is considered the last of the Nordeste Tradition, or if these two dates ("6000 BP" and "5000 BP") refer to the painting and material traditions independently.

30. Or 10,000 - 1500 B.C., Pessis, "Pré-história da Região," 67.

31. Guidon, "Ocupações pré-históricas," 48.

32. Ibid., 72.

33. Guidon, "Seqüencia cultural," 142.

34. Ogel-Ros, "La notion," 67-68.

35. Behling, "High Resolution Holocene," 265-266; Guidon, "Seqüência cultural," 142-143; Kipnis, "Early hunter-gatherers," 586-587; and, Ledru et. al., "Last 50,000 Years," 239-240.

36. Guidon, "Seqüência cultural," 142-143. Another date of 5780 B.C. was also retrieved from material at this site, and was also considered an "Agreste" date (FUMDHAM database).

37. Guidon, "Notas sobre dois sítios," 42-43. See also, Aguiar, "Tradições e estilos," and "A Tradição Agreste."

38. 2000 B.C. in Guidon (1986), "Seqüencia cultural," 143. More recently the proposed terminal date of the Agreste paintings has migrated: 2000-1000 B.C. in Guidon (1998), "Ocupações pré-históricas," 48; and c. 1 B.C. in Pessis (1999), "Chronology and evolution," 47. The most recent proposal (1999; Pessis, "Pré-história da Região,") does not even address the termination of the Agreste Tradition.

39. See the earlier discussion in the Introduction. This has implications for the dating of the Serra Branca Style, which was based upon the same methodology. Serra Branca Style and Salitre Subtradition figures are occasionally painted over earlier styles, but the reverse also occurs. See, for example, the site of Viaduto I discussed earlier in this dissertation (Chapter 2).

40. Geib wrote, "it has yet to be demonstrated that a point style equals a people, and there are many examples of point styles spreading spreading rapidly between different cultural groups" ("Radiocarbon Record," 124).

41. Behling, "High Resolution Holocene," 266; Ledru et. al., "Last 50,000 Years," 239; and, Schmitz, "Prehistoric Brazilian," 74.

42. Behling, "High Resolution Holocene," 253, 266. The current millenium has been the wettest period of the Holocene. However, the last 140 years of colonial impact upon the environment has noticeably diminished the availability of water and created areas of heavily stressed vegetation.

43. Guidon, "Ocupações pré-históricas," 52.

44. Guidon and Arnaud, "Chronology of the New World," 175.

45. Guidon, "Ocupações pré-históricas," 51. Prous reports earlier cultivation of corn and cabaça farther south in Minas Gerais, between 2000 and 1000 B.C. (Prous, "Agricultores," 346).

Additional Excerpts:

Sample Illustrations

Table of Contents

Excerpt from the Introduction

Excerpt from Chapter 1

Excerpt from Chapter 2

Excerpt from Chapter 3

Excerpt from Chapter 4

Excerpt from Conclusions

Excerpt from the Appendix

This site is developed and maintained by Reinaldo Morales Jr., Ph.D.
Department of Art History,
Virginia Commonwealth University, Richmond, Virginia, USA.
Last updated on 6 July 2002. All material © Reinaldo Morales, Jr.
unless otherwise noted. Send comments to
Author takes full responsibility for all content.
This is a web site, not a refereed journal.