PERGAMON Quaternary International 53/54 (1999)21 41

Archaeology of the Pleistocene-Holocene boundary in Brazil

Andr6 Prous a, Emilio Fogaqa b
aFederal University of Minas Gerais State and CNPq-Brazil, French Archaeological Mission of Minas Gerais. Brazil
b Setor de Arqueologia, Federal University of Minas Gerais State. Museu de Hist6ria Natural UFMG,
Cx. P. 1275, 31080-010 Belo Horizonte-MG, Brazil

Abstract

After a short history of Brazilian research about Pleistocene peopling, the Late Pleistocene/Early Holocene changes of climate
and resources are discussed. Claims for a very ancient presence of Man before 13,000BP (Itaborai, Toca da Esperanqa, Pedra
Furada and Lapa Vermelha sites) are polemic, because putative structures and artifacts could have been made by natural processes. In
central and northeastern Brazil some sites dated between 14,000 and 10,000 BP provide clear evidence of human activities and
coexistence with huge Pleistocene mammals.
In a synthesis of our present knowledge about the 12,000/8000 BP period, biological aspects of the Lagoa Santa people and their
subsistence are discussed. The main regional traditions are Umbu with bifacial projectile points, and Humaita with heavy implements
in the, southern part of Brazil. In the central and northeastern regions, the Itaparica tradition is characterized by typical retouched
flakes and limaces. Although projectile points are generally thought to be absent in the earliest levels of this tradition, some have been
found in Boquete Cave. In Amazonia, some cultural aspects seem to have been the same as in central Brazil. (t~..1998 INQUA/Elsevier
Science Ltd. All rights reserved.

1. Introduction In all these localities the dates, the stratigraphy, or the

As in the rest of the Americas, there are great dis-
cussions about when Man arrived in Brazil. The geo-
graphic and climatic features were different enough from
the present conditions and this point makes the work of
the archaeologists difficult. The Late Pleistocene coastal
margin, a probable way of penetration from the north, is
now submerged, and the sites in which Man would have
exploited the rich marine resources until 7000 BP, are no
longer available. The inland sediments were generally
eroded during the Early Holocene, which seems to have
been much wetter than the Late Pleistocene. Only in
places such as caves, where the sedimentary setting might
protect Pleistocene layers from erosion, can ancient sites
be preserved. Pleistocene findings older than 200,000 BP
are claimed at Itaborai and Toca da Esperanqa (Lumley
et al., 1988). Dates between 13,000 and 50,000 years BP
were obtained in Goifis (Barbosa, 1991a), Mato Grosso
(Santa Elina Shelter: Vialou et al., 1995), Minas Gerais
(Lapa Vermelha: Laming-Emperaire, 1979; Prous, 1986a),
and mainly Piaui (Guidon and Arnaud, 1991; Parenti,
1993).
putative artifacts and charcoal considered of anthro-
pogenic origin, are problematic (Fig. 2). Although these
provocative affirmations are challenged, several open air
sites and shelters yielded less polemic radiocarbon dates
between 12,000 and 10,000 BP in the states of Rio
Grande do Sul, Minas Gerais, Mato Grosso, Parfi and
Piaui. Some of these sites provide satisfactory evidences
of human activity. Although not all of them have the
same credibility, it seems that Man was really present in
central Brazil around 11,000 or 12,000 years ago: Espe-
cially in Minas Gerais, rich occupation levels of this
period are found.
For the next two millennia (10,000/8000 BP) there
are many more sites in every state where archaeologists
have worked intensively in the last decade (Rio
Grande do Sul, Minas Gerais, Goifis, Bahia, Pernam-
buco, Rio Grande do Norte, Piaui, Mato Grosso,
Par/0, but the publications are mainly short preliminary
reports. Even for this period, it is impossible to
make a good synthesis of the human occupation in
Brazil, as most of the territory remains archaeologically
unknown.

1040-6182/99/$ see front matter (~" 1998INQUA/ElsevierScience Ltd. All rights reserved.
PII: S 1040-6 18 2(98100005-6
22 A. Prou,s, E. f)~ga'a Qualernal 3' lnlernational 53,54 (1999) 21 41
As we have been asked to discuss the 14,000/8000 BP
period, we shall not discuss the claims of older human
occupations. We include an appendix of references on
sites older than 14,000 BP (Appendix 1).
2. Results and discussion
2. I. Late Pleistocene~Early Holocene climate and natural
FesoIAFCCS
It is difficult to generalize about the past climatic and
ecological changes of a large area such as Brazil, since we
only have information from a few recent studies. The first
attempt to elaborate synthetic models appeared in the
1970s (Ab'Saber, 1977) but these must now be modified
using recent information from the southern, the central
northeastern, and the Amazonian regions.
Today, southern Brazil is characterized by cool and
wet highlands. The upper topographic areas are covered
with parklands and the adjacent lower areas with mixed
forests in which the Araucaria pine provides rich seed
crops. Along the valleys of the Paraguay and Uruguay
rivers, a riverine subtropical forest grows. Highlands also
occupy a great part of central Brazil, covered by the
tropical forest in the eastern wetter area and an arboreal
savannah called the cerrado in the westernmost part,
where the dry season exceeds 3 months. Northeastern
Brazil, with a very long dry season, is mainly covered by
scrub, bush and cactus vegetation called the caatinga.
The north is mostly occupied by the Amazonian equato-
rial forest, but some areas of cerrado are found near
Guyana and Venezuela (Fig. 1).
Between 20/18,000 and 13/11,000 BP, evidence indi-
cates a generally drier climate. In Carajis and northern
Amazonia, savannahs partially replaced the rainforest
until 13,000 years BP (Markgraf and Bradbury, 1982t. In
Lagoa Santa, the intensive formation of carbonate con-
cretions at the caves, indicates that not enough water was
present to dissolve the carbonates. The palynological
record shows that Podocarpus trees disappeared ca. 9300
BP.
In the 1970s, Ab'Saber (1977) stated that this arid
phase was characterized by the expansion of the caatinqa
into central Brazil and the expansion of the cerrados into
Amazonia, where only some forest refuges would have
remained. In these refuges, insulation would have
improved animal speciation processes (Brown and
Ab'Saber, 1979). According to Ochsenius' opinion (19851,
eolian sand accumulations would prove the existence of
true deserts such as "Paragua~u" in northeastern Brazil.
Today, many paleontologists and paleobotanists
(Gu6rin, 1991; Oliveira, 1992; Cartelle, 1995)believe that
the Late Pleistocene faunal remains and palynological
records suggest a vegetation mosaic with forests along
the rivers (even in Northeastern Brazil, that would
have been wetter than nowadays) and various types of
cerrados in other areas.
Generally, dry periods have to be understood as more
seasonal (i.e. long dry winter) rather than reflecting deser-
tic conditions. Thus, the caatinga formations would have
been of very limited extension, since all the animal species
that are living in the caatinga (like Kerodon or Toly-
peutes) also live in the cerrado and have no specific
adaptation to dry climates. The caatinga modern exten-
sion would be a recent phenomenon. In Amazonia, the
forest decrease would have been less important than the
zoologists thought considering the now challenged
refuge theory (Oliveira, 1992; Salgado-Labouriau, 1994).
With the decrease of temperatures and the increase of
open vegetation territories at the end of the Pleistocene,
austral varieties of great mammals, typical of the
Lujanian faunal assemblages of Argentina, migrated
northward. Lestodon (a Giant Sloth) arrived in southern
Brazil (Silo Paulo State), and many others reached much
farther areas, such as central and northeastern Brazil.
Myocastor coypus, Hippidion and Arctotherium, for
instance, were clearly cold adapted animals. Further-
more, Cartelle (1995) recently found bones of Lama
.quanicoe in Toca da Boa Vista (Bahia State). All these
animals met the indigenous tropical mammals, mainly
large armadillos (Pampatherium, Panoctus and Hop-
lophorus), giant Sloths (Eremotherium, Nothrotherium and
Scelidotherium), proboscidian (Haplomastodon waringi),
Litopternae (Xenorhinotherium), Notoungulae
(Tri.qonodops) and the great monkey Brachyteles.
The presence of this last animal demonstrates that the
Atlantic type forest was present in central Bahia, today
a semi-desert region.
At the very end of the Pleistocene, the pine (Araucaria)
forest also grew northward, as the cold period became
wetter. Previously, the very dry conditions did not let this
forest grow into Silo Paulo or Minas Gerais States, as
Ab'Saber (1977) stated.
By 12/10,000 years BP, the climate became wetter and
warmer. In some places, lacustrine and fluvial sediments
reveal important erosive phases at the beginning. Only
later, mostly after 8000 BP, a general expansion of the
torest occurred, as pollen assemblages clearly indicate.
Thus, several geomorphologists think that the Pleisto-
cene/Holocene boundary is older in tropical lowlands
than in temperature regions, where the adjacent ice
sheets caused temperatures to remain lower for two more
millenia (Coltrinari, 1992; H. Kohler, pers. commun.).
It is still uncertain if there were short dry and cold
episodes ca. 13,000 BP or 10,400 BP. In any case, all the
huge animals disappeared in paleontological sites by
10,000/9000 BP. Hippidion returned to Argentina, where
it survived a little longer. Some very recent, unpublished
radiocarbon dates for giant sloth and mastodon from
Bahia suggest that these animals also may have survived
later into the Holocene, but the dates have been obtained
 A. Prous, E. Fogaa / Quaternary International 53/54 (1999) 21-41 23

L *: . : o
.~ r

,'~ANJRO

e~UTH
AMERICA
i0L','
tO0 km 0 6 0 0 km
| 4 i

HOLOCENE VEOETATION
Grol$1ondl ....... Slote boundory
Coollngo M I N A ] OER&I$ Stole
Cerrodoll
Seoso,'~l roreit 0 > 20,000 8P
1;~,,~A,,,o.o.ioh 'fo,o,i ,~ I 0 , 0 0 0 BP
Atlontic forest 0 B,O00/I0,000 BP
IT Arou.tr,o .,, ,o.,., O Mode'ito sHe~
S~o ml~, veoeloliod Potinologicol onolyslS
~ Coo~tal vegetation

REF.: A T L A S GEOGR~FICO, MINISTE'RIO E,A F D U C A ~ , O , FA E, RIO DE Jt~NEIRO, 1990 DES .i MARCO, S BRITO, 1995

Fig. 1. Map of vegetation and sites. 1: Mirador, Parelhas. 2: Lapa do Gavi~o, Carajfis. 3: Madeira river sites. 4, 5: S~o Raimundo Nonato (Pedra
Furada, Sitio do Meio, Baix~_o do Perna, AntoniO.o, Toca do Paraguaio). 6: Central (Toca da Esperanqa, Toca da Onqa, Manuel Lat~o). 7: Abrigo do
Sol. 8: Santa Elina. 9: Serra Geral. 10: Barreiro. 11: Lapa do Drag~o-Montalvtmia. 12: Janu~iria/Peruaqu Valley, Lapa do Boquete. 13: Unai, Lapa do
Gentio. 14: Caiap6nia. 15: Serran6polis, GO JA 01.16: Serra do Cip6, Santana do Riacho. 17: Lagoa Santa (Caieras, Lapa Vermelha; Cerca Grande,
Confins, Sumidouro). 20: Alice BoEr. 21: Itaborai. 22: Uruguay phase sites.
24 4. Prolt~s, E. /.'oeal'~ Quaternary hllernaliona/ 53/54 (1999) 21 41

kl(]. 2

0- -0
:.%,......

f ":-~ , ...". .... "~ ~d

0 ~

~F
0-% !

/_.,:-

~-~-0
iJ }
0- 0
FIG. 2'

Fig. 2. Flaked stones from Pleistocene layer~, a.b: l'oca da t{spcran~l If,onl Lumlc~ et ~d., 1988). c.d: Pedra Vurada ffrom Guidon, 1991). e i: lbicui
phase (e "'ltaqui complex"L from Miller (1971 ). k: Lapa Vermelha. J- I: Santana inferior (from Prous. 19921. 2'. Industry from Lapa do Gavifio Cave.
 A. Prous, E. Fogaqa / Quaternary International 53/54 (1999) 21-41
by new techniques and need to be confirmed (L. Martin,
pers. commun.).
As in other parts of America, the reasons of the extinc-
tion of huge mammals are not clear. There are very
few evidences of killing by Man. Exceptions are an
Eremotherium humerus from Toca da Boa Vista which
has been, without any doubt, disjointed and butchered,
and an iliac bone of Haplomastodon from Lagoa Santa,
possibly showing human cuts (Prous, 1986b). In any case,
the appearance of this new predator, whose hunting
techniques the animals did not know, might have been
a central factor in their extinction. More important, per-
haps, was the decrease of grasslands and the increase of
forest territories. Large Savanna mammals remained iso-
lated and genetically weak while animals such as deers,
tapirs and pecari, were able to survive in the forest. In
some places of central Bahia or Piaui, the beginning of
the desertification during the Early Holocene may be one
of the extinction causes in this region, as herds of giant
sloths are found in holes and caves that were the last
places where permanent water remained available. It is
important to notice that all the large gregarious mam-
mals disappeared, which prevented Man from becoming
a specialized big game hunter.
Lastly, from 17,000 to 7000 BP, Man lost the use of
most of the coastal plains in central and southern Brazil
because of the sea level rise. Owing to the absence of
recent tectonic uplift in Brazil, archaeologists cannot
examine the Pleistocene sites which must have been nu-
merous along the shoreline, certainly an important way
of colonization during the Pleistocene.
2.2. 13,000/10,000 interval
2.2.1. Lapa Vermelha IV (Minas Gerais)
At this shelter of Lagoa Santa region, the skull and
some bones of a woman were deposited in the intermedi-
ate levels between 10,200_+220 BP and 11,680_+ 500
14C years BP. Recently, W. Neves (pers. commun.) ob-
tained a radiocarbon date on these human bones: of
9330 _+60 BP (Beta 84439). However, the organic residue
may come from humic acids, so the real age of the
skeleton might be older. At more recent levels, some
bones and coprolites of a giant Sloth (Scelidotherium) are
dated ca. 9580 _+200 BP (Gif 3208). There were no fire-
places or the typical industry of this period. Thus, dating
was generally performed on scattered charcoal, probably
of natural origin (at this site, the senior author could
prove the presence of natural tree burning in Holocene
sediments; surely this phenomenon also occurred during
the Ple~tocene).
2.2.2. Santa Elina (Mato Grosso)
This cave is being excavated by A. and D. Vialou
(Vialou and Vialou, 1994; Vialou et al., 1995). Under an
25
anthropic pavement dated at 6040 BP is unit II of fluvial
origin, which is dated between 7010 _+ 70 to 10,120 _+ 60
14C BP (Gif 9369 and 3954). At the base of unit II,
numerous dermal bones and some bone fragments of
a giant Sloth (Glossotherium) have been recovered near
a fireplace, associated with flakes and retouched dol-
omitic limestone splinters.
2.2.3. Touro Passos/Lageado dos Fosseis
(Rio Grande do Sul)
Pleistocene sediments were deposited all along the
lower reach of the Uruguay River under a volcanic ash
layer dated at 10,400 _+ 110 14C years BP. On the eroded
terraces, E. Miller found several bones of huge mammals
including Glossotherium robustus (one of them dated at
12,270 _ 220 BP). At the same stratigraphic levels, crude
basalt and quartzite choppers and flakes were found
associated with the rolled bones (Miller, 1974; Rodrigues,
1992). However, after S. Milder (1994) reexamined the
stratigraphy, the lithic material, the field notes and dis-
cussed the results of the site with Miller's assistant
(Jacobus, pets. commun.), many contradictions came out.
Both the artifacts and their association with the bones
(which were found around 4.5 km away from the
14C sample) are doubtful. Finally, Miller's publications
contradict his own field notes. Therefore, the evidences of
human occupations at these ancient layers are not accep-
ted by some archaeologists of Rio Grande do Sul.
2.2.4. Alice Bo~r (S~to Paulo)
This open air site located on a fluvial terrace was first
excavated by Becker (later Beltrfio) in 1964 (Beltrfio,
1974) and more recently by Heredia and Beltr~o (unpub-
lished) with a team that included a geomorphologist.
Part of the industry has been studied by Cunha (1992 and
1994). A rich flint industry was found at unit III which is
a thick sandy layer (excavated by the arbitrary levels
method), that seems to be a secondary deposit; conse-
quently, there were no archaeological structures. The
radiocarbon dates come from floated charcoals while
nine TL dates were performed on burnt flints. The middle
part of unit III yielded ages of about 6000 years BP.
However, two TL dates were recently obtained from the
lower part: 10,950 __%1020 and 10,970 + 1620 BP. In some
way, this confirms a previous date of 14,200 _+ 1150 BP
(SI 1208) (Beltr~to, 1974) with a comment that suggests
"the utmost caution in dealing with SI 1208, for the heavy
dilution.., the already large error quoted might be
trebled". This lower part has the richest industry: mainly
side scrapers, but also notched flakes, some typical
end scrapers and burins. There are pendunculated or
leaf shaped biracial points, some of them with a spiral
body, resulting from special alternating retouches.
Blades and bladelets are not uncommon. The senior
author, who examined part of this material in 1971, was
very impressed by the high quality of some retouched
26 ,4. Prous, E. k%ga~'a / Quaternap3 International 53/54 (1999) 21 41
artifacts and also the occurrence of true burins, unusual
instruments in South American prehistory.
At the deeper stratigraphic unit V, a gravel fluvial
deposit, Beltrfio found crude flakes and chopper-like flint
chunks that she interpreted to be of human origin. How-
ever, according to the senior author of this paper, Tixier
(pers. commun.) and Bryan (1978), this material is the
result of natural processes, with the exception of a typical
but unique end scraper, whose original location is not
clear.
Problems relative to the sedimentation process, m-
consistency between some dates and the stratigraphic
sequence, and lack of information about the field pro-
cedures, make use of this site difficult, even if an Early
Holocene or Late Pleistocene occupation seems quite
likely.
2.2.5. Abrigo do Sol (Mato Grosso)
This cave in the Amazonian region was excavated by
E. Miller in the 1970s, but we only know of short papers
by Miller (1987) and Puttkamer (1979). A crude and
atypical cobble industry is present at the lower levels
(down to 5.4m depth), dated between 12,000 and 9000
BP in square N 11/12, and 14,500 BP in square O 11/12.
However, the stratigraphic location of this industry re-
mains unknown, probably due to stratigraphic distur-
bance. Hence, no reliable information on Pleistocene
Man is available from this site.
2.2.6. Cerca Grande (Minas Gerais)
W. Hurt and O. Blasi worked at several shelters in this
locality of the Lagoa Santa region. Shelter VI, with 20 m 2
excavated, provided two radiocarbon dates for the lower-
most metre: 10,378 _+ 122 BP, 9720 _+ 128 BP. From this
period, there are several graves of Layoa Santa Man,
numerous quartz flakes, some end scrapers, bifacial pro-
jectile points and ground axes (Hurt and Blasi, 19691 that
characterize Hurt and Blasi's Cerca Gramle complex.
They did not see any typological modification between
ancient and recent occupations, even those with ceram-
ics. Perusing this material at the Rio de Janeiro National
Museum, the senior author observed that nearly all of the
material was flaked by the bipolar technique, and most of
the quartz "scrapers" were, in fact, bipolar cores (Prous,
in Leroi-Gourhan, 1988), although some true flint
scrapers exist. Human remains have been described by
Alvim (1977) and are now being studied by Neves.
2.2.7. GO JA Ol (Goihs)
The lower levels of this quartzite cave are radiocarbon
dated from 9060_+ 65 (S1 3698) to 10,740_+ 90 BP (N
2348) and characterize the Paranaiba phase of south-
western Goifis. Big quartzite flakes unifacially retouched
are typical of this phase. Numerous end and side
scrapers, thick double side scrapers with one pointed end
(limaces) are present. Awl and polished spatulas were
made with deer bones. Some beads made with bones or
egg shells have been found. Faunal remains mostly in-
clude armadillos, lizards, and snails. Some deer, large
rodents and a few fish have also been found (Moreira,
1981; Schmitz et al., 1976).
2.2.8. Other findings older than lO, O00 BP
In several other localities, ancient dates have been
obtained, but we will only briefly describe these findings.
Some of these sites have not been published, and others
have recently been widely divulgated (Pedra Pintada).
The Boquete and Santana sites will be detailed in the
next section of this paper.
2.2.9. Lapa Pintada CaL'e (Pard)
A first report about this Amazonian cave near Monte
Alegre has been recently published (Roosevelt et al.,
1996). The lowermost strata provided many consistent
radiocarbon from ca. 11,200 to 10,000 BP. Many flakes
and some retouched instruments (among them limaces
and bifacial points) were found. Iron pigments were nu-
merous and may have been used to paint the walls of the
cave, as few pigments were recovered in the Holocene
levels. Subsistence remains include fruits available during
the rainy season, diverse mammals, reptiles, mussels and
fish, suggesting an opportunistic strategy.
2.2.10. Barreiro site (Goihs)
F'rom an unpublished dissertation (Martins, 1993), we
know that level 5 of this open air site, located at a terrace
near Rio Cocal, has been radiocarbon dated to 10,605
BP. The senior author examined part of the associated
lithic collection which was mainly composed of chal-
cedony and quartz flakes, some of them chipped by the
bipolar technique, and very few retouched artifacts.
2.2. I I. Gentio 11 site (Minas Gerais)
In this cave, excavated by O. Dias, three radiocarbon
dates are known from the lowest levels, ranging from
10,190_+120 (SI 6837) to 8 6 2 0 1 1 0 B P (Gif 3208)
tSchmitz, 19871. Secondary burials with burnt human
bones are described by Machado (1990).
2.2.12. Drag~o Shelter (Minas Gerais)
The lower levels (8 and 10) are radiocarbon dated
between 10,000_+255 and 11,000+300 years BP
(CDTN, 1007 and CDTN, 1008, with s. 95%). The chert
and sandstone industry, still unpublished, includes some
thick unifacially retouched quartzite instruments. Chal-
cedony and fine small and medium size flakes of chert are
typical of the thinning or retouching process of delicate
artifacts. Many of them show reddish colours resulting
from accidental thermal processes. No hearth structures
of the oldest periods have been found, and the dates come
from isolated charcoal pieces (Prous et al., 1984).
 A. Prous, E. Fogaga/ Quaternary International 53/54 (1999) 21-41
2.2.13. Boquete cave (Minas Gerais)
This cave yielded various occupation levels (units VII
and VIII), well dated between 12,000+ 300 ~4C and
9870 + 260 years BP (CDTN 1084 and CDTN 1077).
These levels have a rich lithic industry, bone instruments,
pit structures, hearths, food remains and red pigments.
A big engraved block with pecked figures, and polished
grooves and cupules was discovered in 1995, covered
by unit VI and a fireplace dated at 9350+-80 BP
(Cal BC 8450 to 8265-Beta 98573). A more full descrip-
tion of the lowest level is given in the last section of this
paper.
2.2.14. Santana do Riacho Shelter (Minas Gerais)
This shelter has provided a charcoal concentration
dated circa 18,000 BP at unit 8. As there were no asso-
ciated human artifacts or structures, we believe that this
charcoal concentration must have been the result of
a natural fire. At unit 7, stratigraphically higher, a thick
layer of ash and charcoal has been partially eroded by
subsequent water action under more humid climatic
conditions. We found some small, simple crystal
flakes and processed ochre fragments, along with a
little stone covered with pigments (natural or anthropic?).
The charcoal was radiocarbon dated at 11,960 _+250 BP
(Gif 5089). However, we have to admit that though
we have carefully excavated the site in order to
identify every stratigraphic disturbance, the pigments
and flakes we found at the non-eroded location,
might be intrusive from the 9700/8400 year old
graves excavated in the upper unit 4. A small animal may
have used one of these burials to excavate deeper along
the rock, and reach the remaining part of level 7 into
which some small flakes and pigments could have fallen
down.
2.2.15. Pernambuco State
Two poorly known sites yielded very old dates
associated with a flake industry: Brejo da Madre de
Deus rockshelter [11,060+90 BP, see Schmitz
(1987) and Lima (1991)-1, and the ChSo do Caboclo
open site (11,000_ 250 BP, MC 1046 - - see Laroche,
1975).
As we have seen, only few scholars believe that archae-
ologists have found sites older than 12,000 BP. However,
no one in Brazil would deny the possibility of Man's
entrance before this period.
Some findings around 12,000/10,000 BP are dubious,
but many sites of this interval show unquestionable occu-
pations. Thus, there is no doubt that Man was present in
central Brazil and Amazonia shortly before ca. 12,000
BP. Evidence is less clear in the southern part of Brazil,
where there were fewer systematic excavations. More
recent sites, between 10,000 and 8000 BP, are much more
numerous and less controversial, and thus they will not
be described here.
27
3. Synthesis of the first secure occupations
(13,000/8000 BP)
We will now review our present knowledge about Late
Pleistocene and Early Holocene Man and Nature. How-
ever, it is necessary to consider that paleoecological re-
search was very uncommon prior to the late 1980s, and
that archaeological excavations, still limited, are mainly
test pits excavated by artificial levels that have been only
partially described in publications. Generalizations are,
thus, very hazardous.
Human bones remains. There are no dated human
bones older than 13,000 BP. Mostly, the Material of the
13,000/8000 period comes from the Lagoa Santa/Serra
do Cip6 region in Minas Gerais (Lapa Vermelha,
Sumidouro II, Lapa Mortuaria near Confins, Escrivania
III, Cerca Grande VI, and Santana do Riacho sites) and
all have the typical Lagoa Santa Man morphology. Most
of these human remains (from nearly 200 individuals) are
deposited in Copenhagen (Zoological Museum), Rio
de Janeiro (Museu Nacional) and Belo Horizonte
(Museu de Hist6ria Natural). As described by many
authors (recently Alvim, 1977, 1992; Neves and
Pucciarelli, 1991; Soto Heim, 1994), the Lagoa Santa type
is very homogeneous and shows characteristic features.
Of submedium size, these people had a moderate sex-
ual dimorphism, gracile bones and long forearms. The
skull was extremely dolichocephalic, with a lengthened
occipital part and occipital bun, a sagittal keeling,
parieto-mastoidian depression under the parietal hump;
broad and short face and nose, rectangular orbit. Non-
metric characters (such as oleocranian perforation, 3rd
molar absence), also attest a great homogeneity sugges-
ting genetic isolation.
Pathological evidence has been studied by Ferrigolo
(1987) and Souza (1992). They show a high frequency of
minor traumas among younger people. Osteomyelite is
also frequent, sometimes intense, and so are arthroses in
temporal/mandibular articulation (due either to hard
vegetal mastication or from technological use of the
teeth). Decayed teeth are not rare. Molar eruption is
extraordinarily late in comparison with other tooth and
bone development (Radicchi, in Souza, 1992).
Subsistence. For the Paleoindian/early Archaic period,
information about subsistence comes exclusively from
shelters. No killing sites are known, which is not surpris-
ing, since the huge animals which had to be butchered in
the same place where they were killed no longer existed.
The non gregarious tapirs were the only big mammals
that remained.
In southern Brazil, only the Afonso Rodrigues shelter
provided bones of medium size mammals such as deer
(Ribeiro et al., 1989, and pers. commun.). However, for the
early period, there were mainly gastropods shells and
eggs of Rhea americana. This material has not been pub-
lished in detail.
28 .4. Prous, E, I'bgaCa Qualernal 3" hlternationa/ 53:54 (1999) 21 41
In central and northeastern Brazil, where cerrados
provide plenty of fruits during the rainy season but
medium size animals are scarce, the evidence from some
Goifis (GO.JA.01), Minas Gerais (Santana do Riacho,
Boquete) and Pernambuco (Brejo da Madre de Deusl
shelters suggest that prehistoric Man was a non special-
ized hunter of some medium size animals (deers and
armadillos), and even smaller prey such as big lizards and
rodents. Apparently, the use of these small animals be-
came more and more frequent after 9000 or 8000 BP.
Fish and bird remains exist, but are very rare: egg shells
occur at Brejo da Madre de Deus. In Goi'as and Pernam-
buco, pieces of wax suggest honey collection, still an
important gathering and ritual activity today. Fresh-
water mussels seem to have been an important part of the
diet in Boquete ca. 10,000 BP. Later, however, they
virtually disappeared, replaced by Me qalohulimus.
a giant land snail which occurred at this cave and
elsewhere until recent times. In Lapa Vermelha after
8000 BP, these wet season snails were the only faunal
remains in many of the small cooking hearths left by
small bands that visited the shelter overnight during the
rainy season.
Dental observations Lagoa Santa Man show a great
frequency of decayed teeth, which in turn suggests great
importance of vegetal food. In spite of vegetal fragility.
findings of carbonized palm nuts (mainly Acrocomia,
Astrocaryum, Cocos coronata, Syattrius} are common
at all the shelters. Fruits of pequi, a typical cerrado
tree (Caryocar brasilienset very rich in vitamins, and
Hymenaea fruit capsules also appear in some of the
Santana graves (Prous and Malta, Eds, 1991). In south-
ern Brazil, Arecastrum palm nuts have been found at
Afonso Rodrigues shelter.
Usually, faunal studies only list taxa and number of
elements; no analysis of taphonomic processes, spatial
distribution or butchering techniques is performed.
Therefore, it is difficult to know if part of the recorded
animals were really prey or only cave dwellers. Thus, we
have little doubt that the faunal remains in the shelters
do not give an accurate testimony of subsistence. In some
layers of the large excavations at Boquete and Santana.
most of the deer bones we found were paws which had
been carried to the site to make bone spatulas.
3.1. Regional traditions
Sites are rarely described, so archaeological structures or
rituals cannot be used to compare cultures. Conse-
quently, the archaeologists usually define prehistoric cul-
tures mainly from the lithic industry, as it is the best
preserved and described material. Nevertheless, few cul-
tural units have been proposed in Brazilian archaeology
due to the following reasons.
Obvious chronological or spatial typological differ-
ences between lithic complexes seem to be scarce. There
are very few qualitative and quantitative studies of collec-
tions; lithic technology was barely studied until a few
years ago. Consequently, in the identification of lithic
units, archaeologists use, almost exclusively, few typolog-
ical criteria, (i.e. presence/absence of some instrument
such as projectile points (southern Brazil) or limaees
(central Brazil), a preference for thin flakes or heavy
instruments.
3.1. I. Southern Bra=ilian traditions (Fig. 3)
In spite of some contradictions between Miller's
(1969, 1976), Kern's (1982, 1991) Ribeiro's (1990) and
Rodrigues' (1992) publications, we will try to present
a general framework of human occupations before 8000
BP in Rio Grande do Sul State.
3.1.1.1. Phases with prq/ectile points. If accepted, the
most ancient archaeological sites are those of Ibicui
(2 sites, radiocarbon dated to 12,770_+220, SI 801:
12,690_+ 100, SI 2351) and Uruguai (10,400_+ 110 and
9595_+ 175 14C years BP) phases in the lower valley of
the Uruguay river. Most of them have been found in
fluvial terraces now covered by the forest and laterally
eroded by waterflow. The sites are situated near islands
or rapids (good places for fishing).
al During the Ibicui phase, vegetation was probably
scarce, the river less important, and Man may have
hunted GIossotherium and Glyptodons, whose bones
are present in the same geological deposits. No projec-
tile points have been found and the lithic material
is simple, sometimes crudely retouched flakes and
choppers.
bl During the Uruguai phase, there were small bifacial
stemmed chalcedony projectile points with triangular
shapes, which were pressure retouched. Side and
end scrapers, various retouched flakes and basalt
chopping tools also were present, and pounders
and anvilstones were frequent. In the Laranjito site,
excavated by Miller, lithic instruments and carbon-
ized fruits formed clusters around hearths (Schmitz,
1987).
c) Recently, two shelters with the same lithic industry
have been excavated. Schmitz established the undated
and unpublished Capivara phase with the material he
found in one of the shelters, and Ribeiro (1989) the
Batinga phase with the lower layers (160/230 cm) of
the Afonso Rodrigues site. At this shelter, the
170/180cm artificial layer is dated to 8290_+130
1"~C years BP (Beta 32183). Nearly 3000 medium and
small metaquartzite flakes were recovered, along with
7 projectile points, some bifacial blanks and one deer
antler used for retouch. No scrapers appeared; a flake
from the wall (?) was stained with red pigment. Also
some bones, egg shells, palm nuts and stones from
hearths were found.
 A. Prous, E. Fogaa / Quaternary" International 53/54 (1999) 21-41 29

\ g

~) ~.tl~

Fig. 3. Early Holocenelayers southerntraditions, a c: Batinga phase (from Ribeiro et al., 1989). More recent material f j: Umbu phase (from
Prous, 1992). k o: Humaitfi Tradition, Altoparanaense industry (from Kern, 1991).

Uruguai, Batinga and Capivara phases would be the
ancestors of the Umbu Tradition, typical of the open land
hunters during the following millenia, characterized by
its carefully made projectile points, scrapers and other
retouched small or medium size flakes.
d) Three fishtail projectile points have been found in
Santa Catarina and Paranfi States, but none in
a dated context. The senior author collected a fishtail
projectile point in a coastal sand dune (Prous, 1992),
where it may have been probably brought by more
recent sambaqui dwellers. Another projectile point
was found at an eroded site near Itapiranga (Rohr,
1968). The only fishtail projectile point recovered
from a stratigraphic context appeared in the undated
Vinitu site (Chmyz, 1978), probably ca. 6000 years old.
3.1.1.2. Occupations without projectile points. Along the
Parana river, Rohr (1968) excavated the Itapiranga terra-
ces and collected heavy basalt artifacts. On the upper
6 meters, dated at 7260_+ 100 14C years BP (SI 440), he
found a typical AItoparanaense industry, characterized by
big flakes and very thick bifacies, some of them
long and straight, others boomerang-shaped which were
similar to those described by O. Menghin in Misiones,
Argentina (Menghin, 1957). This material has been
30 A. Prous, E. Fogaqa / Quaternarv International 53/54 (1999) 21 41
studied by Schmitz and Becker (1968). At a lower depth
(7 m), Rohr found similar flakes - - but no bifacies .... with
a hearth dated at 8640 _+95 14C years BP (SI 995), that he
interpreted of altoparanaense origin. For Brazilian ar-
chaeologists, the Altoparanense industry belongs to the
HumaitLt Tradition that occupied exclusively riverine for-
ests along the lower reaches of main rivers. Humaita
people are supposed to have been mainly gatherers, per-
haps also fishermen, with an industry characterized by
heavy basalt or metaquartzite instruments such as chop-
ping tools and thick scrapers; they did not use stone
projectile points.
Thus, in the southern region, Umbu and Humaita
traditions would have been coeval, from ca. 8000 BP.
Umbu's and Humaitfi's northern boundary seems to be
the Tropic of Capricorn, in the state of S~o Paulo, where
no sites older than 7000 BP are known. It is impossible to
say if there were some contacts between the southern
traditions and the cultures of central Brazil.
3.2. Central and northeastern archaeological units
Many shelters occupied in limestone and sandstone
formations in Central and Northeastern Brazil (states of
Bahia, Goifis, Mato Grosso, Minas Gerais, Pernambuco,
Piaui and Rio Grande do Norte) provide a great amount
of information. However, it is difficult to make compari-
sons between the different regions studied.
3.2.1. Central Minas Gerias
No dated ancient open air sites are known from this
period, but many shelters were occasionally used by Man
until 8000 years BP. In the Lagoa Santa region they used
to face karstic lakes in transitional regions between the
forest and the cerrados. In Serra do Cip6, prehistoric
men occupied shelters near brooks and torrents at
the boundary between the cerrados and the scarce herba-
ceous mountain vegetation. In both regions, vegetation
resources were important during the summer period
where quartz, mainly crystal, was the available raw
material.
Several shelters were used as cemeteries (among others,
Confins, Cerca Grande, Sumidouro, Santana do Riacho),
while isolated human skeletons or bones came from
other places (Caieras II, Lapa Vermelha). The bodies
were in graves covered with stones, some of them with
red pigments.
During this period, quartz was worked by bipolar
technique on anvilstones, so we found many thousands of
small simple flakes. Larger flakes (more than 2 cm),
which were frequently extracted by free hand percussion,
were retouched into end scrapers. Microwear analysis
indicates that they were mainly used to fix ochre powder.
Quartzite slabs, where available, were retouched in side
scrapers for wood working. A few retouched small flakes
and a broken flint limace showed that this imported
material was a precious one. Pebbles were used as anvil-
stones to chip quartz, but also to break palm nuts.
Greenstones were brought from kilometers away and
ironstone (hematite) from long distances to make ground
axes. Polished instruments seem to be very ancient in this
region: axe fragments and blanks have been found in
layers dated from 10,000 to 8000 years BP in Cerca
Grande (Hurt and Blasi, 1969) and Santana do Riacho
(Prous and Malta, 1991). Very scarce projectile points
and thinning small flakes have been found in the same
site. If the excavations were less extensive, it would have
been easy to think that this period was a "pre-projectile
point" one.
Few bone tools were found, the most interesting of
which is a possible hook fragment from Santana do
Riacho and some spatula fragments, bone and antler
points. Vegetal beads and rope fragments from ham-
mocks were also found in the same site.
Even though thousands of paintings cover the walls of
many shelters, we do not know if they were made during
this period. In Santana, however, some scrapers were
stained with yellow powder that had not been used in the
graves; perhaps, these pigments may have been made as
early as ca. 9000 years BP to draw on the walls. However,
at the moment, the oldest dated rock paintings in
Santana and Lapa Vermelha, are between 7000/5000
and 4000 BP (animalistic Planalto Tradition).
3.2.2. Northern Minas Gerais
In the Januaria and Montalvfinia regions, several shel-
ters with old dates are being excavated. The older layers
of Lapa do Boquete and Lapa do Dragfio have been
dated from 12,000 to 11,000 years BP. There, silex, chal-
cedony and metamorphic sandstone extracted from the
highest hills of the region, were the main raw materials to
be chipped. Sandstone was selected to make the heavy
coarse limaces and thick scrapers up to 10 cm in length,
which are characteristic of the period before 8000
years BP in this region (Fig. 4) and in the neighbouring
Goi~ts State. Microwear analysis indicates that these in-
struments were used to work wood (Prous et al., 1992).
Bone industry is poor, mainly having polished spatulas
made of metapodial deer bones.
The excavations at Boquete cave revealed a great
fallen boulder covered with grooved petroglyphs and
buried by 9000/7000 year old years. Some red pigments
have been found in the oldest layers, but we do not know
if they were used for the oldest paintings of the cave walls
(SCto Francisco geometric Tradition). After ca. 9000
years BP, the occupation remains in both this and nearby
sites became completely different.
3.2.3. Goihs
The oldest dates come from the southwestern part of
the State, near Serran6polis, where 40 m 2 were excavated
by Schmitz (1989) in GO-JA-01 shelter. A rich industry
 A. Prous, E. Fogaqa/ Quaternary International 53/54 (1999) 21-41 31

....

Fig. 4. Industryfromlowerunit (8), Lapa do Boquete

was found in the lower levels. Characteristic of this
period were large, frequently more than 10 cm thick,
unifacially retouched scrapers and limaces made of
quartzite blades extracted from the walls, and very small
(upto 3 cm long) flint cores. There also are bone spatulas
and antler instruments.
Subsistence remains suggest a diversified habitat and
unspecialized hunting and gathering. This archaeological
complex, dated between 10,740+90 (N 2348) and
8370__ 85 (SI 5562) in 8 shelters, is known as the Para-
naiba phase. Later, the characteristics of human occupa-
tions changed completely.
Similar lithic material, dated at 10,750+ 300
14C years BP was found in GO-NI-49 shelter, located
100 km north of Serran6polis. Here, artifacts were mostly
made of quartz. Two quartz unifacial projectile points
were associated with the typical scrapers (Barbosa et al.,
1976), but the archaeologists fear the points may have
been intrusive.
3.2.4. Northeastern region: Bahia, Pernambuco,
Rio Grande do Norte and Piaui states
In Pil~o limestone shelter (central Bahia), A. Bryan
and R. Gruhn (1993) excavated a 35 cm thick layer that
provided four 14C dates between 8790 _ 80 (Beta 10014)
and 9650 _ 90 years BP (Beta 10015) with some crystal,
quartzite, chert (brought from many kilometers away)
and limestone material. There were quite exclusively
small (2 to 4 cm long) simple flakes; "some of them with
edges steppened by retouched and heavey use", but also
32 A. Prous, E. t:oj~a{'a QuaternaJT hlternalional 53:54 (1999) 21 41
some quartzite scrapers, limestone choppers, pebble
hammerstones and few chert core fragments. Quartz vein
was used after being flaked by the bipolar technique.
Freshwater mussel shells (Diplodon) were used as
scrapers; also one pointed deer bone was present. Char-
coal and ash were abundant and so were faunal remains
of medium size animals such as deer, peccary and arma-
dillo. Few scattered human bones were recovered. As
other sites of the region, Pilfio shelter was abandoned
during the Middle Holocene.
Furna do Estrago (Pernambuco) excavated by Lima
(19911 stands near one of the rare permanent water points
in the region. The lowest level, characterized by charcoal
and ash lenses, was dated between 8495_+70 and
9150_+ 1#C years BP (SI6296 and 6298) but has been
poorly described. Deep hearths in which many food
remains were left, suggested a permanent settlement.
Most of the bones were from small animals such as
rodents and armadillos. Some egg shells and propolis
(honeybee products are of great importance in all in-
digenous cultures) also appeared. Plant remains were
scarce; including Spondia, Sya,qrius palm nuts, and resina
of Piptadenia, a medicinal plant. Small quartz and flint
flakes were the only lithic material found. After ca.
8000 BP, the site was abandoned.
In the same state, Gruta do Padre ('ave (Itaparica
region, Pernambuco) provided rich lithic material.
mainly thick scrapers and planes made of pebbles
collected at the nearby Silo Francisco river. Based on
this material, Schmitz created his ltaparica Tradition
in which he included all the Early Holocene industries
characterized by thick unifacial instruments (especially
limaces) and, theoretically, without biracial projectile
points.
Mirador site (Rio Grande do Norte), is a very small
rockshelter (12 m2). Some burials, mainly of children
with beads, recovered from the lowest level t 60 cm,
dated at 9640_+ 100 ~'~C years BP CSIC 720) would be
"in contact with the lowest paintings of the pannel'" (Luft,
1989). According to G. Martin (1991). this would indicate
an age of ca. 10,000 years for the rock art. belonging to
the Serido style of Nordeste Tradition, which depicts
small anthropomorphic figures in hunting and ritualistic
scenes.
In the Silo Raimundo Nonato region (Piaui), climatic
instability would have characterized the Early Holocene
(wetter phase 10,000/9000 BP: increasing aridity since
9000 BP). The industries 14,300_+ 400 and 12,200_+ 60 BP
in Toca do Sitio do Meio until 8450 years BP in Pedra
Furada constituted the Serra Talhada phase. Though
quartzite choppers like the Pleistocene ones continued to
appear, the lithic material of this period is mainly of flint.
chalcedony and quartz. In general, there were flakes.
some of which were unifacially retouched (end and side
scrapers, notched flakes). Most carefully worked artifacts
(limaces, borers) were shaped from large chalcedony
flakes. Heat treatment is considered to have been used
for manufacturing of the quartzite double scraper type
(Parenti, 1993). At Pedra Furada, a quartz bifacial
projectile point has been found and is thought to be the
result of barter, and not of local manufacturing. Recently,
at Toca do Meio, a ceramic sherd would have been
recently recovered from a hearth radiocarbon dated at
8960 years BP. A polished stone was found at the lower
level, dated at 9200 BP (Guidon, 1994/5).
At Toca do Baixfio do Perna shelter, various rock
paintings, depicting deer, nandu, a wild cat and small
anthropomorphic figures (some of them buried before
the excavations), stand just aboce a 9700 BP layer and
under a radiocarbon dated level of 4920_+70 BP (Gif
7739, Martin 1996); Guidon (1989) thinks they are at
least 9700 years old because people would not have sat to
paint, but we disagree with this interpretation. Also,
following Guidon, the Nordeste Rock Art Tradition
(hunting or ritualistic scenes carefully painted) and
A.qreste Tradition (with coarse great figures) coexisted in
the Silo Raimundo Nonato region at the end of the Serra
Talhada phase.
3.2.5. lVorkshops and quarries
Thousands of cores, cortical flakes and some finished
Paranaiba like instruments were found along quartzite or
flint outcrops at quarries which appeared in several open
air sites of central Brazil (Goifis, Minas Gerais, Bahia).
Although these sites remain undated, it seems they were
probably worked since the Early Holocene.
Near Niquel'andia (Goi/ls) several workshops were
found on fluvial terraces. There, quartz and chalcedony
pebbles (the senior author found many bipolar products
in the collections with clear evidences of percussion with
a wood o1" antler hammer in some of them), flint and
metamorphic sandstone were flaked by prehistoric Men.
Few retouched instruments such as typical limaces ap-
pear at the Barreiro site where the lower layer V provided
a 10,605 BP date (D. Martin, 1993). In Goias state, the
junior author found hundreds of Itaparica like retouched
artifacts, dehita,qe waste and an in situ associated pro-
jectile points workshop; they were under 2 m of Early
Holocene sediments, in a fluvial terrace of the Tocantins
River (Fig. 5a d).
Concluding, we can say that most of the Early
Pleistocene levels in central and northeastern Brazil are
characterized by an industry with large thick flakes
(Fig. 5), which are frequently unifacially retouched
(limace type is the most representative, being most care-
fully worked in Goifis than in other states). Bone spatulas
are also typical while projectile points tend to be absent.
Apparent exceptions may be explained by the specificity
of certain raw materials, mainly quartz. Large flakes
cannot be obtained from medium or small size crystals.
It is clear that Lagoa Santa and Santana's industries
could not show thick huge scrapers and limaees. For this
 A. Prous, E. Fogafa / Quaternary International 53/54 (1999) 21 41 33

(4) '

! i

Fig. 5. ltaparica industry a-d: Tocantins project (Serra da Mesa, Goi/ts). e-i: Gruta do Padre (from Martin and Roacha, 1990). j: Pedra Furada
(1bid.). k--n: CaiapSnia (from Schmitz et al, 1986). o: deer bone spatula.

reason, in these sites the few medium size artifacts of
quartz vein, unifacially retouched, are significant. Likely,
they are the equivalent of the largest flint or sandstone
instruments of the Itaparica Tradition of northern regions.
Another important point is the absence of projectile
points. We saw that some appeared in Goi~is and Piaui,
where they were considered intrusive. This argument
does not apply in Lapa do Boquete, where we identified
the manufacturing waste. Otherwise, what would be the
unknown culture responsible for the biracial projectile
points, which does not appear in the archaeological re-
cords? We think that Itaparica groups did make the
points, but they left them at the shelters only when the
points were broken. If archaeologists peruse the .refuse
that they classify as "microblade", they will probably find
the proof of point manufacturing. Unlike southern
hunters, ltaparica people did not easily leave their most
carefully made artifacts.
3.3. Northern and western regions
In the western Pantanal swamps, no preceramic site
has been found so far. In the higher western lands, Santa
Elina shelter (Vialou et al., 1995) will provide important
information, but the results of the recent excavation are
mostly unpublished. The industry seems to be very crude
(marginally retouched limestone slabs), similar to the
putative Pleistocene.
34 .'1. Proud, E. Foga~'a Quaternap?' International 53:54 (1999) 2l 4l
In Amazonia, Lapa do Gavifio has been studied by
Lopes and Hilbert (Silveira, 1994); 40 of the 350m z were
excavated in the 1980s. Thousands of quartz and am-
ethyst crystal flaked by bipolar technics (Fig. 2at were
found in level A2, dated from 7900_+45 (Geochron Labs)
to 8140_+ 130 BP (Teledyne Isotopes). The raw materials
were brought from at least 30 km away. Inside the cave,
cores and flakes were left along the walls: outside, on
a sheltered area, lithic material surrounded a big dis-
turbed hearth with faunal remains.
4. Spatial organization of archaeological remains
at Santana do Riacho rockshelter and Boquete cave,
Minas Gerais Santana do Riacho Rockshelter
Santana do Riacho shelter, which stands at the l\)ot of
the Serra do Cipo Mountains, can be seen from a long
distance. Thus, it must have been a territorial reference in
several periods, as it is suggested by the thousands of
paintings on its walls. Between ca. 9700 and 8230 BP, the
site was occupied by the Lagoa Santa people who took
advantage of the two well separated cave platforms. The
upper platform was used as a cemetery, mainly during
the 8500/8200 BP period. On the lower platform, lithic
material, mainly reflecting the daily life, was recovered:
including thousands of small flakes of local quartz that
were later swept along the wall. Carefully retouched
crystal scrapers and borers were concentrated in a 15 m 2
surface where fallen stones were less frequent. There,
worn out scrapers made on quartzite slabs were re-
touched, and clusters of little splinters revealed the place
where this occurred. These retouched scrapers were re-
used to work wood, as indicated by microwear analysis,
until they broke and the fragments, some refitted at the
laboratory, thrown away. More precious flint instru-
ments were also retouched. However, since they were
carefully carried away, we only found small splinters.
Raw material used in the preparation of yellow, orange
and dark red mineral pigments and some stained lithic
flakes were found along the wall. A small boulder with
one face painted, suggested that rock painting may have
begun in this period.
On the upper platform, the cemetery (Fig. 6) stands in
a 15 m 2, apse-like natural spot whose center was occu-
pied by a huge fallen slab buried by sediment accumula-
tion after 8000 BP. Prehistoric people dug the gritty soil,
sometimes making a box with some slabs and the bent
corpses were put down in the graves. In several se-
pultures we found rope fragments. In one burial, the rope
fragments seemed to belong to a hammock where
a young person laid with an infant wrapped in plant
fibers upon his pelvis. A hearth was lit near the pit and
the live coals mixed with the finer sediments, were thrown
into the grave, heating or burning some bones. Finally,
they built a cairn with the blocks, some of them painted.
During the latest period, ca. 8300 BP, there were no more
free places around the great fallen slab, and some burials
were partly destroyed in order to bury new corpses. The
disturbed ancient bones were then superficially buried
again among the stones of the cairn, sometimes with
ochre deposit. Three burials of adult females, far from the
huge slab, might have been either the latest, or those of
less important or "different" people as the lack of
red powder at their grave suggested. One of these corpses
was a woman with cranial pathology and probably with
mental problems (Neves, pers. commun.).
In the cemetery, both men and women were present,
but women died a bit younger. Little children were also
numerous and frequently received a greater amount of
concentrated red pigments and collars of vegetal beads.
Paleopathological studies made by S. Souza (1992)
showed several inflammatory bone processes and Harris
lines, which point to frequent periods of physical stress.
The cemetery was used during both wet and dry seasons.
We found carbonized Caryoear seeds with shells of giant
snails at some graves while in others, seeds only available
in dry seasons such as those of Pterodon pubescens. The
lithic material found at the undisturbed graves was
mostly composed of bipolar flakes and crude scrapers for
pigment (usually stained with yellow, though pigments of
this colour seldom appear in the graves). The more deli-
cate quartz artifacts or specialized quartzite side scrapers,
common in the lower platform, were almost completely
absent.
Santana do Riacho shelter is, at this date, the best
known and published site in Brazil. It documents various
aspects of prehistoric life. Unfortunately, we did not find
the spaces where they cooked and discarded food re-
mains. Perhaps, there were unexcavated parts of the site,
but we think that prehistoric men probably occupied
nearby places that we have not found yet.
4. 1. Lapa do Boquete Cave
The Boquete Cave is situated close to the little Peruaqu
River, in northern Minas Gerais. There, more than 60
sites, mainly shelters with rock art, have been recently
discovered in an area of about 10kin 2. The very dry
conditions at the cave preserved the organic material.
The excavations revealed the existence of nine main geo-
logical units which were subdivided into 27 archaeologi-
cal layers. We excavated about 20% of the cave in five
places, and now we are working in the open air occupa-
tion, out of the sheltered area. We shall now present the
lower archaeological level of excavation no. 1 (25 m2),
dated from 12,000 to 11,000 BP.
This excavation was made in the western part of the
cave, along a huge fallen stalagtite that constituted a kind
of wall. Preliminary analysis of some clusters of material
seems to represent different types of activities and may
have been coeval (Fig. 7). The clay sediment of unit VIII,
 A. Prous, E. Fogaga / Quaternary International 53/54 (1999) 21-41 35

91JI

4-

mx

rl 1 73

T -I- Q ~ I ~ ' * L WaLL

4~ L O ~ ANDSLABS

~J
~- F

~lO,m I
W.q~I

ltllI

Fig. 6. Santana do Riacho ancient cemetery,upper platform

frequently mixed with organic material, is aggregated in
the zones where the more humid climatic conditions of
the Early Holocene let the water percolate. Archaeologi-
cal remains (down to more or less 1 cm 2) and structures
were mapped; and the smallest material recovered in the
sieves was localized in its own square meter map. In this
way, it is possible to study the first occupation floor of
the region (Ribeiro et al., 1996).
4.1.1. Making, using and discarding lithic instruments
A big anvilstone used to break nuts and a ball of red
prepared pigments occupied the center of an "empty"
circular area in JK/11-12; an accumulation of flint waste
was around this cleaned area. There were shaping and
retouching flakes from scraper manufacturing, thinning
flakes from projectile points and one bifacial point frag-
ment. Three combustion structures (charcoal and ash
36 .4. Pr()u.s, E. /~'og(~(t Quctlerna 0 lnternation(d 53,54 (1999) 21 41

PROJECTILE POINTS
~o~: FLINTWORKiNG WASTE
r" T T "I coRE
RETO~JCHED ARTEFACT
~) HAMMERSTONE
1~ ANVIL STONE
M
LIMESTONE SLA8S
,,~" FRESH WATER SHELLS
%~. PALM NUTS
T~ MAMMAL BONES
O HEARTH
POST HOLE
O HOLE
CONCRETION

" tB; 2' '

J
lm
L I

J
r

k
+ +

"~ ~, H
P lPv ~O
I. . "_M P
~ L .L _t.'- % *-'Na.- ". J. _J

8 9 10 12
11

l-ig. 7. First o c c u p a t i o n lloor in B o q u c t c C a v e . e x c a v a t i o n no. 1

lenses) were located on the boundary of this circle. In the them had been used to scrape dry wood (wood was not
southern structure, a carefully shaped scraper and some preserved in this layer, but post holes were found at the
expedient tools (irregularly retouched flakes) had been limit of this area). Though some discarded instruments
discarded. Microwear analysis revealed that some of were thrown in active fireplaces which is suggested by the
 A. Prous, E. Fogaqa / Quaterna~ International 53/54 (1999) 21-41
thermal alterations they display, retouched artifacts were
generally left in another place, along the western calcite
cemented wall.
4.1.2. Hearths
Most of the subsistence remains were found near or
within the hearths; great concentrations of carbonized
palm nuts broken upon an anvilstone were found at K10,
J10 and L9, situation similar to those at K l l and M8,
which showed typical pecked circular depressions. Palm
nut wood, rich in oil, provides excellent coal. Other
plants, like Cnidosculus and Sterculia seeds, also ap-
peared near some hearths, mainly in JL9 and JK10. More
than half of the bone refuse (836 g) concentrates in the
hearths around JK/ll-12 structure. These are mostly
from small deer (Mazama): teeth, ribs and paw fragments
were all probably rejected during artifact manufacturing.
Scutes of armadillo (Euphractus and Dasypus) were pos-
sibly used for cooking purposes. Bones of little animals
including birds, fish, and tortoises were found in JKL/8-9
and may be associated with the J10 hearth. Near the L10
fireplace, we found thousands of half calcinated bivalvia
shells. In addition, there were some isolated small fresh-
water molluscs (Pomacea and Limnea), probably brought
to the site with water or some other aquatic supply, as
they are so minute that Men were not able to perceive
them.
4.1.3. M8 structure
This structure stands in a poorly lit part of the cave.
A deep artificial cylindrical hole contained anvilstones,
few bones and chipping refuse. A hammerstone, lime-
stone splinters and several retouched instruments were
left at the margin of this hole. Within a short distance,
nearly all the silicified sandstone artifacts of unit VIII
were located. Was there a deposit to store this valuable
material (the raw material was available some kilometers
away and shaped at the quarry), later used as a refuse
area?
It is clear that the first stages of the reduction process
were absent at Boquete's ancient levels or, at least in the
six excavated areas, as virtually no cortex appeared on
the flakes. Unretouched large flakes were also absent,
and very few cores were collected (Prous et al., 1992;
Foga~a and Lima, 1991; Foga~a, in press). Thus, these
operations were made in other sites. Butchering was not
a local process, unless the entire bones were carefully
swept. Such a cleaning process seems to have existed, as
in JK/11-12. Likely, the flaking refuse was first discarded
in the hearths and only later pushed away, shaping the
circle described.
The destruction of non carbonized plant remains in
lower units prevents us from better knowing gathering
strategies, although the occurrence of palm nuts suggests
a summer occupation. Another question is if all the
clusters represent one or several occupations. This latter
37
hypothesis would fit better with some observations, but
we have to conclude our analysis before we try to answer
that question.
In spite of these constraints, the excavation of large
surfaces and the detailed analysis of minute lithic refuse
at Boquete permitted us to recognize two important
points. First, artifacts not present at the shelters have
been carried away, although they have been shaped at
these places. Some of them have been found in the open
air site, 20m from the shelter, where complete artifacts
were left, and neither shaping nor retouching activities
occurred. Second, some aspects of daily life during the
Prehistory have been clarified.
5. Conclusion
Ancient or Middle Pleistocene human occupation
is claimed in Brazil (Lumley et al., 1988), but the evidence
is very weak (Prous, 1995). Upper Pleistocene Man
is supported by Pedra Furada, Santa Elina and Lapa
Vermelha's lowest levels but, in the latter shelter, little
evidence is available prior to 11,960 BP. Santa Elina's
lower levels are only known from previous information
about faunal remains. At Pedra Furada, Guidon and
Parenti's interpretations are challenged, although few
scholars deny the possibility that some of the chipped
stones and burnt places might have an anthropic origin.
During the Early Holocene, no fluted point or leaf
shaped projectile point Horizon is known. However, the
obvious presence of Man between 12,000 and 11,000 BP
in Minas Gerais and Piaui States points to a more
ancient arrival in Amazonian or coastal regions. Littoral
- - now submerged - - sites are not available, and Ama-
zonia is poorly known since it is very difficult to explore
and few suitable localities with preserved Late Pleisto-
cene sediments are present. However, Lapa Pintada,
Lapa do Sol and Gavifio shelters prove that sites with
a good stratigraphic sequence do exist.
If really there were human occupations during the
Middle and Late Pleistocene before 12,000 BP and
people left their marks in some of the sites we described,
we have to recognize that they used very little stone
which is quite possible in regions where wood was
very abundant (in some recent protohistoric sites, the
chipped stones of Tupiguarani Indians, for instance, are
not much more characteristic). Also, they would have
had no interest in selecting good raw material and in-
stead used only bad local ones. Low density and high
mobility, typical of pioneer people, would explain the
scarcity of sites and incontestable evidences. Besides this,
the fact that hardly a "conservative" archaeologist would
test a putatively sterile Pleistocene sediment reinforces
this situation.
In the face of the present "evidences", the Faithful
and the Unbelievers will have useless discussions, as the
38 A. Prous, E. k'oga;'a Quaternm 3, International 53,'54 11999) 2l 41
arguments of each depend on subjective interpretation of
the facts. However, it is necessary that Unbelievers con-
tinue criticizing controversial data (if not, "Pleistocene
sites" will multiply out of control ...) and it is also impor-
tant that faithful archaeologists remain working in
Pleistocene sediments. If they do not, possibly more re-
presentative ancient sites will never appear.
By 12,000/10,000 years ago, the central and north-
eastern regions were inhabited by at least one type of
a scarce, non mongolized and genetically isolated people:
the Lagoa Santa Man. Sometimes, they would have oc-
cupied rockshelters to work and to bury people, painted
some walls and engraved fallen rocks by 10,000./8000 BP.
Unfortunately, open air sites have not been found. Dur-
ing this interval, lithic remains were characterized by the
use of local and exotic raw materials at the shelters. They
produced large thick flakes, frequently unifacially re-
touched; side and end scrapers, planes and limaces, with
some regional variations, were typical, in spite of Braz-
ilian archaeologists' opinions, some evidence points to
the production of bifacial projectile points: the other
characteristics of the so called ltaparica tradition are not
fully expressed. Nor is the chronological significance of
the presence of famous limaces, since this artifact also
appears in much more recent sites or layers.
Though chronological data and some isolated findings
show that the pioneers met huge mammals like giant
sloths, subsistence refuse and site location suggest un-
specialized hunters/gatherers.
By 8000 BP, everything abruptly changed in central
Brazil. The shelters were much less occupied and fre-
quently completely abandoned. Large flakes and re-
touched artifacts disappeared though scarce microflakes
still can be found at the shelters. Shells of giant snails
efficiently substituted stone scrapers to work wood.
Gastropods were also much more exploited and the ratio
of small animals increased in faunal refuse. We also
believe that S?to Francisco and Planalto rock art tradi-
tions appeared at this time. Did the new inhabitants
whose physical morphology proves that they were differ-
ent from the Lagoa Santa Man, at least in Peruaqu
Valley, bring a new way of life and discarded the ancient
technologies?
Barbosa (199 lb), according to Greenberg's ideas and
glottochronological data, suggests that by 10,000 BP the
extension of the Amazonian Forest (at the expense of
cerrado) would have been responsible for the separation
between proto G& Pano and Karib languages. Each
group would have migrated to a peripheral region, with
G6 indian ancestors in the new cerrados that were replac-
ing Pleistocene caatinya in Brazil. Thus, they would have
been the origin of the Itaparica Tradition. Though this is
an attractive hypothesis, we have to remember that
ltaparica like industries are present in Minas Gerais
before 10,000 BP. We previously pointed out that
recent research suggests that mixed cerrados formations
remained in central Brazil during the Pleistocene, and
that the Amazonian forest did not decrease in such
a drastic manner as it was thought some years ago.
Otherwise, the hypothesis of Barbosa would fit chrono-
logically better for the arrival of new populations by 8000
BP in the states of Minas and Goifis.
Archaeologically little is known for this period in the
state of Silo Paulo, which is considered the boundary
between the central industries and the southern tradi-
tions. Would this boundary mean that Early Holocene
cultures respected exactly the modern ecological regions
and that their lithic industries had to be completely
dependent on strict adaptation to climate and vegetation?
We know very little to make such deterministic hypothesis.
In the southern states, since ca. 7000 BP, industries are
also divided into two main groups, the Umbu and
Humaita traditions, but it is not clear if the most ancient
industries known in Rio Grande do Sul would be their
ancestors. Projectile points have been used as a diagnos-
tic/ossile-directeur for the Umbu tradition and would
exist in the Uruguay River Valley since the Late Pleisto-
cene. Theoretically, the points found until ca. 7000 BP
would have a triangular shape different from the more
recent typical early Umbu points, that would generally be
leaf-shaped (Ribeiro, 1990). But this point is challenged
by A. Schmidt (commun. in SAB Congress, 1995). Other-
wise, Early Holocene industries lack some carefully re-
touched artifacts such as the more recent Umbu scrapers.
In any case, and in contrast to central Brazil, there is no
decline in stone chipping technology and there is no
abrupt change; only a clear opposition between riverine
and upland industries that is thought to reflect cultural
and non seasonal diversity.
As the Lagoa Santa Man along with the Tequendama
Man, is one of the two oldest known American popula-
tions, anthropologists discuss his origins, believing he
would be representative of the primeval inhabitants of
the Americas.
As there are no results available on DNA analysis
lpreliminary tests are being made at the Laboratory of
Molecular Biology of Minas Gerais University), anthro-
pologists are using mainly multivariate analysis of metric
characters for their comparative studies with other mod-
ern or archaeological populations. Neves and Pucciarelli
(1991), from 13 craniometric variables, suggest that
Lagoa Santa Man would be much more like paleo-
australian or upper Zhoukoudian cave dwellers than like
any other population. They think that a pre mongoloid
population of northeastern Asia would be at the origin
of at least two migratory waves: one southward (to
Australia) and the other northward (to Beringia and
Americas), Lagoa Santa Man being representative of the
latter. Soto Heim (1994) believes that the broad face
would be the sign of an incipient mongolization, which
does not exist in other paleoindians skulls such as those
of Chinchorro. It is not clear if Soto Heim thinks that
 A. Prous, E. Fogaga/ Quaternary International 53/54 (1999) 21-41
such phenomenon would be a sign that Lagoa Santa
Man ancestors would have entered America in a second
migratory wave, or if she believes it would be the result
of a regional evolutionary process that occurred in
America. In any case, Lagoa Santa Man seems to have
disappeared by 8000 or 7000 BP in the Lagoa Santa
region. Perhaps it is only because Men did not bury dead
people at the shelters anymore that we do not have
skeletons of the later Archaic period. But in other areas,
including the state of Minas Gerais, the few and generally
unpublished findings suggest that a new population ar-
rived and occupied the previous settlements.
Acknowledgements
We thank Irmhild Wrist (pers. commun, about Goi~ts
sites and bibliography), D. Meltzer and an anonymous
reviewer for their comments and suggestions on an early
draft of this manuscript, and Marcos Eug6nio Brito (map
and illustrations).
Appendix
References on sites and conditions older than 14,000 BP
Bigarella, J., Beltr~io, M., Toht, E., 1984. Registro de fauna na arte
rupestre: possiveis implicaq6es geol6gicas. Revista de arqueologia,
Bel6m 2, 31-37.
Bombin, M., Bryan, A., 1978. New perspectives on early Man in
Southwestern Rio Grande do Sul, Brazil. In: Bryan, A. (Ed.), Early
Man in America, from a Circum Pacific Perspective, Occasional
Papers, Department of Anthropology, University of Alberta 1, pp.
301-302.
Bryan, A., 1991. The fluted point Tradition in the Americas - - One of
several adaptations to Late Pleistocene American Environments.
In: Bonnischen, Turnmire (Eds), Clovis: origins and adaptations,
Center for the Study of first Americans, Oregon State University,
Corvallis, pp. 15-33.
Cann, R., 1994. Mt DNA and Native Americans: a Southern Perspect-
ive. American Journal of Human Genetic 55, 7-11.
Espinet-Moucadel, J., Beltrgo, M.C. da C., 1987. Etude trac6ologique
de deux pieces lithique de la Toca de Esperanqa, R+gion de Central,
l~tat de Bahia, Br6sil. L'Anthropologie 91,943-948, Paris.
Gamb6ri, L., 1991. O Pleistoceno na ~trea arqueol6gica de Sao
Raimundo Nonato, Piaui, Brasil. Clio, s6rie arqueologia, Recife 4,
21 22.
Gomes, M.N.S., 1987. L'industrie lithique et la stratigraphie du site
d'Itaborai (Rio de Janeiro, Br6sil). M6moire de DEA, Museum
d'Histoire Naturelle, Paris 49 p.
Guidon, N., Delibrias, G., 1986. Carbon 14 dates point to Man in the
Americas before 32,000 years ago. Nature 320, 769-771.
Hurt, W., 1960. The cultural complexes from the Lagoa Santa Region,
American Anthropologist 62, 569 585.
Kohler, H., Pil6, L.B., 1991. The Quaternary Chronology of the mor-
phogenetic events in the karstic region of Lagoa Santa/MG, Brazil.
XIII INQUA Congress, Beijing, Abstracts 1, 177.
Lage, M.C., 1990. t~tude arch6om+trique de l'art rupestre du Sud-est du
Piaui. Ph.D. Dissertation, Paris 407 p.
Ledru, M-P., 1992. Etude de la pluie pollinique actuelle des for&s du
Br6sil Central. Ph.D. Dissertation, IPH, Paris.
39
Meltzer, D.J., Adovasio, J.M., Dillehay, T.D., 1995. On a Pleistocene
human occupation at Pedra Furada, Brazil, Antiquity 68, 695 714.
Miller, E., 1992. Adaptaq~o agricola pr6-hist6rica no Alto Rio Madeira.
In: Meggers, B. (Ed.), Prehistoria de Sulamerica - - Nuevas perspec-
tivas, pp. 219-229.
Ochsenius, C., 1983. Aridity and biogeography in Northernmost South
America during the Late Pleistocene. Zbl. Geol.a~ Palgontologie,
Stuttgart, Teil 1(3/4), 264-278.
Peyre, E., 1993. Une am+ricaine de 10.000 ans. Archeologia, Dijon 295,
60-61.
Prous, A., 1990. Le Grand Abri de Santana do Riacho. Les Dossiers
d'Arch6ologie, Dijon 145, 74-77.
Prous, A., 1991. Les fouilles de la Lapa do Boquete. Journal Soci6t6
Am6ricanistes, Paris 77, 77-109.
Prous, A., 1992. Arqueologia Brasileira, Universidade de Brasilia ed
(UNB), Brasilia 605 p.
Prous, A., 1993. Pedra Furada: mais sobre a chegada do Homem nas
Am6ricas. Ciencia Hoje, Rio de Janeiro 16, 8 9.
Prous, A., 1994a. L'Arch6ologie br+silienne aujourd'hui. Probl+mes et
tendances. In: L6vfique, P. (Ed.), Recherches br6siliennes, Annales
Litt6raires de l'Universit~ de Besan~on 130, 9~43.
Prous, A., 1994b. L'Art rupestre du Br6sil. Bulletin de la Soci6t6 Pr6his-
torique Ari~ge-Pyrdndes, Tarascon sur Ari6ge 49, 77-144.
Prous, A., Lima, M., Foga~a, E., Brito, M., 1992. A indflstria litica da
camada VIII da Lapa do Boquete, Vale do Perua~u, MG, Brasil,
3 Congreso ABEQUA, Anais, Belo Horizonte 342-362.
Prous, A., Malta, I. (Eds), 1991. "Santana do Riacho" Arquivos Museu
Historia Natural UFMG, Belo Horizonte 12 (1991) and 13 (1992),
384 + 400 p.
Ribeiro, P. Mentz Klamt, P., Buchaim, J., Ribeiro, C., 1989. Levan-
tamentos arqueol6gicos na encosta do Planalto entre o vale dos rios
Taquari e Cai, RS, Brasil. Revista do CEPA 16, 49-89.
Schaffee, S., Hyman, M., Rowe, M., 1993. AMS 14C dating of rock
paintings. In: Steinbring, J., Watchman, A. (Eds), Time and Space,
Dating and Spatial considerations in Rock Art Research, Mel-
bourne, pp. 67-73.
Servant, M., Soubi+s, F., Suguio, K., Turcq, B., Fournier, M., 1989.
Alluvial fans in Southeastern Brazil as Evidence of Early Holocene
dry Climate Period. International Symposium on Global Changes
in South America during the Quaternary, $5.o Paulo, Special Publi-
cations 1, 75-77.
Stewart, T., Walter, H.D., 1955. Fluor analysis of putatively ancient
animal and human bones from Confins Cave, Minas Gerais, Brazil.
31 Congress of Americanists, S~o Paulo, pp. 925-937.
References
Ab'Saber, A., 1977. Espaqos ocupados pela expans~o dos climas secos
na Am6rica do Sul por ocas~.o dos periodos glaciais quatern~trios
Paleoclimas, Universidade de $5_o Paulo 3, 19 p.
Alvim, M., 1977. Os antigos habitantes da ~rea arqueol6gica de Lagoa
Santa, Minas Gerais, Brasil-estudo morfol6gico. Arquivos Museu
Hist6ria. Natural UFMG, Belo Horizonte 2, 119 173.
Alvim, M., 1992. Os antigos habitantes da Serra do Cip6, MG-Brasil
- - estudo morfol6gico peliminar. Arquivos Museu Hist6ria Natural
UFMG, Belo Horizonte 13, 107-128.
Barbosa, A., 1991a. Projeto Serra Geral. Clio, serie arqueologia, Recife
4, 35-38.
Barbosa, A., 1991b. A Tradiqao Itaparica: Uma compreensao ecol6gica
e cultural do povoamento inicial do planalto central brasileiro. In:
Meggers, B. (Ed.), Prehistoria de Sulamerica - - Nuevas perspectivas,
pp. 145 160.
Barbosa, A., Schmitz, P., Miranda, A., 1976/7. Um sitio paleo-indio no
m+dio-norte de Goi~s. Anuario Divulgaq~.o Cientifica, Goi~.nia 3/4,
21-29.
40 1. 1)]'()113', ]']. ]+'()~.~[l~'([ Q u a l e r ,
Beltffto. M., 1974. Datag'6es arqueoldgicas mats antigas do grasil.
Anais Academia Brasileira de Ci,?ncias, Rio de .laneiro 46, 211 251
Bombin. M., 1976. Modem paleo-ecol6gico cvoluti:o para o nee+
quatern/uio da regifio de Campanha. oeste do Rio (irande do Sul.
Brasil. Communicacion. Museu dc Ci~?ncia da PtiC'-RS. Porto
Alegre 15. I 91.
Brown. K., Ab'Sabcr. A.. 1979. Ice forest rel'uge atld exolution in the
nee-tropics: correlation of paleoclimatological, geomoFphological
and pedological data with modern biological endemism, l'alco-
climas, Silo Paulo 5. 30 p.
Bryan, A. {Ed.), 1978. Early Man in America, from a ('ircttm Pacilic
Perspective. Occasional Papers, Department of Anthropolog_~, tlm-
versity of Alberta 1.
Bryan, A.. Gruhn. R.. 1993. Archaeological investigations at six caxes o]
Rockshelters in interior Bahia. Brazil. BrazilIan Studies, t inivesit3 of
oregon, pp, I 166.
Cartelle, C., 1995. A fauna local da Toca da Boa Vista. l)is>ertation.
University of Minas Gerais. Brazil.
Chmyz, 1., 1978. Projeto arqueol6gico Itaipil-3 Relatorio. ('uriliba.
141 p.
Coltrinari. L.. 1992. Paleoambientes Quatcrnarios nit +&md'rica do Sul:
primeira aproxima~;/io. Anais 3 Congrcso dc ABEQt IA. Belo ttori-
zonte, pp. 13 42.
Cunha, L.M., 1992. La region de Rio Clare .... Mice BoOr, I'11.11. I)isscr-
ration, Mused d'Histoire Naturellc, Paris.
Cunha. L.M., 1994. Le site d'Alicc Bo~r IBresil). l,'Anthropologie, Pari,
98. I10 127.
Ferrigolo, J.. 1987. Palcopatologia comparada tie :erlebrados: <)
Homem de Lagoa S i n l t a , o Hotllem dos sambaquis tie ('abegttda>+
c mamiferos pleistoc6nicos, Ph.D. Dissertation. [!I~R(IS, Perle
Alegre, 490 p. + fig.
Foga~;a, E. m press. A tradi~o llaparica c its indt~lslrias l~rc-ccrgtmica~,
da Lapa do Boquetc (Minas Gerais). Revista do Mtlseu tie Arquc-
ologia e Etnologia. Si~o Paulo, USP.
l-ogm;a, E.. Lima. M . 1991. L'abri du Boquctu IBrdsilt: les premiere,
industries lithiques de l'Holocene..lournal Socidtc des Ambricanistc~.
Paris 77. 111 123.
Gu&in, C., 1991. Lit faun de xcrtdbrds du Pleistocdne Supt}rieur de Silo
Raimundo Nonato. Comptes Rendus de I'Acad,Smic des Sciences.
Paris, 312, serie l[, 567 572.
Guidon, N., 1989. Nolas s(;brc dots sition arqueologico:, dc Sgto
Raimundo Nonalo, Piaui, Clio, Recife. ( nivcrsidadc Federal dc
Pernambuco, serie Arqueologia 5, 41 46,
Guidon, N., 1991. Nota sobre dots sltios da (u'ea arqueoldgica de S~'lo
Raimundo Nonato, Piaui. Bresil. Clio. sdrie arqueologia. Reciti: 4,
19 20.
Guidon. N.. 1994 5. Pesquisa> al-qucologicas em Site Raimundo
Nonato. Piaui. Novos dados. Revista dc Arqueoh+gia, S. Paulo 8,
37 46.
Guidon. N., Arnaud, B., 1991. The Chronolog> of Nc\~ World: t\vt+
faces of one reality. World Archaeology 23. 1(+7 178.
Hurt. W.. Blast. O.. 1969. O projeto arqueol6gico Lagoa Santa, Mina,~
Gerais. Brasil. Arquivos Museu Paranaensc. Curitiba, scric Arquc-
ologia 4, 63 p.
Kern, A.. 1982. Le pr;ecdramique du plateau sud-brcsilicn. Ph.l).
Dissertation, Universit;e de Paris, Publicaq:6es A\ulsas. (}PRS.
Universidade Federal de Rio Grande do Sul. Porto Alegre, Brazil 1.
427.
Kern. A., 1991. Arqueologia Prd-Histdrica dc Rio Grande do Sul.
Mercado Abertn, Porto Alegre 356 p.
Laming-Emperaire. A.. 1979. Missions Archcologiques I~ranco
Br+siliennes de Lagoa Santa, Minas Gerais, Br+sil. kc Grand abri de
Lapa Vermelha (PLI. Revista de Pre-historia, Sin+ Paulo I, 53 89.
Laroche. A., 1975. Contribuig6cs para a Arqucologia Permtmbucana.
Col6gio Pernambucano, Recife 52 p.
Leroi-Gourhan, A. (Ed.I, 1988. Dictionnairc dc la Prdhistoirc. Pttt-,
1' ed 1222 p. Second revised cd. 1994. 1277 p. see entries aboul
w 3' l,lc'rmtliona/ 53 54 ( 1 9 9 9 ) 2 1 41
Brazilian sites written by N. Guidon (Pedra Furada) and A. Prous
(Lapa Vermelha. Alice Boer, Boquete, Cerca Grande, Santana do
Riacho, G O JA 0l).
l,irqa, i.D.. 1991. Dots pcriodos de subsistencia no Agreste Pernanl-
butane: 9.000 e 2.000AP. Clio. sdrie arqueologia 4, 57 61.
I+uft. \'.. 1989. Os restos alimentares do stile Mirador, no Boqueirao de
Parelhas, RN, Clio. Rccife 5, 27 32.
l.umley, H., kumley, M., Beltfflo. M., Yokoyama, Y., Labcyrie, J.,
Danon, J.. Delib~ rias. G.. Falguieres, C., Bischoff. J., 1988. Pr&encc
d'outils laillds associes ik une faune datdc du Pleistocene Moyen dans
llt Toca da Esperant;a, rdgion de Central, Etat de Bahia, Br6sil.
l,'Anthropologie, Paris 91, 917 942.
\lachado, L. Cheuiche. 199(I. Sobrc its pfftticas funemrias de crcma~,'/io
c seas \aria6es enl grutas do llorle +2 noroestc de Minas Gerais,
Revista do CEPA, SIa Cruz do Sul 17, 235 247.
\htrkgra['. V., Bradbur}, ,l.P., 1982. Holocene climatic history of South
\ m e r i c a , In: Mangerud. J., Birks, H.J.B., Hager, K.D. (Eds), Chrono-
stratigraphic subdivision of the Holoccne, Striae 16, 40 45.
Martin. (i.. 1991. Novos dados sobre as pinturas rupcstres do Scrid3,
Rio (il'ande do N o n e Clio. seric arqueologia, Rccife 4, 141 145.
Martin. G., 1996. Pre-Hist0ria do Nordeste do Brasil. Universidade
Federal de Pernambuco, Recite 395 p.
\hu-lin, G., Rocha. J.. 1986. Ind0strias liticas em ltaparica, no Vale do
Middle Silo l~rancisco [Pernambuco, Brasil) Clio, Recife 8, sdric
arqueoh)gia 3, 99 135.
Martins, D., 1993. P[analtina de Got(is. unpublished M.A. Dissertation.
/qli~clsit 3 of Silo Patio, Brazil,
\lcnghin, O., 1957. El poblamiento prehistorico dc Misiones, Analcs dc
'\rqueologia ~ Etnologia. Mendoza 12. 19 40.
Mildcl. S.. 1994. A "'Fase Ibicui': urea revisfio arqueol6gica, crono-
logica e estratigr;ifica, unpublished M.A. dissertation PlYC:RS,
P<.;rto Alegrc.
Miller, E.T., I969. Pesquisas anqueologicas cfetuadas no noroeste de
Rio Grandc do Sul (Alto L ruguail In: Programa Nacional de Pcs-
quisas Arqueoldgicas, Beldm 3. 13 32.
Miller. E.F., 1971. Pesquisas arqueol,Sgicas efetuadas no Planalto Mcr-
idional de Rio Grande do Sul, rios Uruguai, Pelotas e das Antas.
P R O N A P A IPrograma Nacional de Pesquisas Arqueoldgicas), rc-
suhados prcliminares do 4 ano, 1968 1969, Belem, Museu Emilio
(ioeldi 4. 37 70.
Miller. E.T.. 1974. Resuhados prelimmares das pesquisas arqueologicas
paleomdigenas c no Rio Grande do Sul, Brasil. Anales del 4t
( o n g r e s o de Americanistas, Mdxico 3, 483 49l.
Miller, E.. 1987. Pesquisas al+queoldgicas paleoindigenas no Brasil
()cidcnlal. In: Nuflez, L., Meggcrs, B. (Eds). Investigaciones
t)alcoindias al sur de la linea cuatorial, gstudios Atacamefios 8,
97 61
Morcira, L.E., 1981. Amilise dos restos de alimentos de origem animal
no programa arqueologico de Goifis. Anuario dc divulga~;/io
cientilica u c a o , Goi/ts 10.98 112.
News. W.. Pucciarclli, H.. 1991. Morphological allinitics of the [irst
Americans: an exploratory analysis based on early Sonth American
human remains. Journal of Htm~an Evolution 21,261 273.
()chsenius. C.. 1985. Pleniglacial Desertization, Large-Atmnal Mass
Extinction and Pleistocene Holocenc Boundary m South America.
Rcxista de Geogralia N o n e Grande, Universidad Catolica tie Chile,
Santiago I2, 35 47.
Oliveira, P. de, 1992. A palynological record of Kate Quaternary veg-
etational and climatic change in Southeastern Brazil. Ph.D. Disserta-
tion. The Ohio State Universityr
Parenti. I-"., 1993. ke gisemenl Quaternaire de la Toca do Boqucir~io da
Pedra F'urada, Piaui, Brdsil. Ph.D. Dissertation, EHESS, Paris 411
p. + lig.
I'rous. A.. 1986a. Os mats antigos vestigios arqueologicos no Brasil
('cnlral. In: Bryan, A. ted.), Early Man in America from a Circum
Pacific Perspective, Occasional papers, Department of Anthropol-
ogy, [:lliversity of Alberta 1, 173 182.
 A. Prous, E. Fogafa / Quaterna~" International 53/54 (1999) 21-41
Prous, A., 1986b. L'Arch6ologie au Br6sil, 300 si6cles d'occupation
humaine. L'Anthropologie, Paris 90, 257-306.
Prous, A., 1992. Arqueologia Brasileira, Universidade de Brasilia ed.
(UNB), Brasilia 605 p.
Prous, A., 1995. Archaeological analysis of the oldest settlements in
America. Brazilian Journal of Genetics 18, 689-699.
Puttkamer, W. yon, 1979. Man in the Amazonian Stone Age present,
meets Stone Age Past, National Geographic 155, 343 389.
Ribeiro, L., Alonso, M., Fogaqa, E., 1996. Produq~o utilizaq~o de
artefatos liticos: uma reconstituiq~_o do espaqo ocupado no inicio do
Holoceno na Lapa do Boquete (Minas Gerais, Brasil). 8 Reuni~,o
SAB, Porto Alegre, Atas 2, 17-30.
Ribeiro, P. Mentz, 1990. A Tradiq~_o Umbu no Sul do Brasil. Revista do
CEPA 20, 129--151.
Rodrigues, J.A., 1992. Arqueologia del Sudeste de Sudam~rica. In:
Meggers, B. (Ed.), Prehistoria de Sudamerica Nuevas perspectivas,
pp. 177 209.
Rohr, J., 1968. Achados arqueol6gicos em Itapiranga. Pesquisas, S~o
Leopoldo, s6rie antropologia 18, 49-51.
Roosevelt, A., Costa, L., Machado, L., Michab, N., Mercier, N., Val-
ladas, H., Feathers, J., Barnett, W., Silveira, M., Henderson, H., Silva,
J., Chernoff, B., Rease, D., Homan, J., Toht, N., Schick, K., 1996.
Paleoindian Cave Dwellers in the Amazon: The peopling of the
Americas. Science 272, 373-384.
Salgado-Labouriau, M.L., 1994. Hist6ria ecol6gica da Terra. Edgar
Bliicher ed., S~.o Paulo 307 p.
Schmitz, P.I., 1987. Prehistoric Hunters and Gatherers of Brazil.
Journal of World Prehistory I, 53-126.
41
Schmitz, P.I., Barbosa, A.S., Jacobus, A., Ribeiro, M.B., 1989. Arque-
ologia nos cerrados do Brasil central, Serran6polis 1. Pesquisas, S'fio
Leopoldo, Instituto Anquietano de Pesquisas, s6rie arqueologia 44,
208 p.
Schmitz, P.I., Barbosa, A.S., Wrist, I., 1976. Arqueologia de Goifis em
1976. mim.
Schmitz, P.I., Becker, I., 1968. Urea indfistria litica de tipo Alto Para-
naense-ltapiranga-SC Pesquisas, s6rie Antropologia, S~.o Leopoldo
18, 21 46.
Silveira, M. Imazio da, 1994. Estudo sobre estrat6gias de sub-
sist6ncia de caqadores-coletores pr~-hist6ricos do sitio Gruta
do Gavi~o, Caraj~,s, PA, S~o Paulo, M.A. unpublished dissertation,
151 p.
Soto-Heim, P., 1994. Les Hommes de Lagoa Santa (Br6sil). Caract6res
anthropologiques et position parmi d'autres populations pal+o-
indiennes d'Am~rique. L'Anthropologie, Paris 98, 81 109.
Souza, S.M. de, 1992. Paleopatologia da popula~_o do Grande Abrigo
de Santana do Riacho, Minas Gerais. Arquivos Museu Hist6ria
Natural UFMG, Belo Horizonte 13, 129-160.
Vialou, A., Vialou, D., 1994. Les premiers peuplements pr6historiques
du Mato Grosso, Bulletin Soci6t6 Pr~historique Franqaise 91,
257 263.
Vialou, A., Aubry, T., Benabdelhadi, M., Cartelle, C., Figuti, L., Fon-
tugne, M., Solari, M.E., Vialou, D., 1995. D6couverte de Mylodon-
tinae dans un habitat pr6historique dat6 du Mato Grosso (Br6sil):
l'abri rupestre Santa Elina. Comptes-Rendus Acad6mie des Sciences,
Paris 320, s~rie IIa, 655 661.