Online chapter

Las sociedades originarias
(The Indigenous societies)

Director
Teresa Rojas Rabiela (Mexico)

Codirector
John V. Murra (United States of America)

Chapter 2: The Original Peopling of Latin America
Alan L. Bryan

Early sites reported in South America will be reviewed by first looking at northern South America, and then traversing the Andes before moving back north from Patagonia to Northeast Brazil. It can be assumed that some coastally adapted people expanding out of the forested Central American funnel turned eastward along the Caribbean coast while other groups advanced southward along the Pacific Coast. So far, little has been reported from eastern Venezuela, the Guianas, or the Amazon Basin, mainly because it is so difficult to find preceramic sites in densely forested regions. Recently, however, two well-stratifed sites have been discovered in the gallery forest along the Rio Orinoco in the Venezuela/Colombia border region (Barse l990). Quartz crystal flakes, a hammerstone, a ground stone axe fragment, and a nut cracking stone were associated with a hearth that dated 9020 + l00 B. P. Other undated preceramic localities have yielded flakes, flake scrapers, and two tanged projectile points in later levels. The utilized flakes and flake scrapers are thought to have been used to manufacture wood, cane, and bone artefacts. A preliminary note (American Antiquity 53: l79) indicates that the site of Taperinha, a large shellmound near the mouth of the Tapajos River before it flows into the Amazon has yielded dates back to l0,000 B. P. Although artefacts are not described, a report on this site will be important because it will describe bone and shell as well as stone artefacts

People expanding eastward from the Darien region of Panama and Colombia soon ran out of dense tropical forests and had to adapt to the semiarid thorn forests of north-eastern Colombia and north-western Venezuela north of the northern terminus of the Andes. The Pedregal River has a series of low terraces and old higher surfaces with good exposures. The Venezuelan archaeologist José Cruxent collected artifacts made of local quartzite from these old surfaces. He discovered a possible technological sequence correlated with the different levels. The highest level above the river yielded large crude cores, scrapers, knives, and rough thick bifaces. The next lower surface yielded smaller, thinner bifaces in addition to the early types. On terrace l, thick willow-leaf-shaped projectile points were added to older types. The thick biconvex points, obviously shaped to fit into socketed hafts like a cylindrical bone or wood point, were termed El Jobo after the nearly country store. Finally, the floodplain terrace yielded small tanged points with constricted bases (Rouse and Cruxent l963, fig. 5 and plate 3). The hypothetical topological sequence, which might be expected if knappers were experimenting with stone as a useful material for tipping wooden spears, remains undated. Acting upon his hunch that the El Jobo points were used to hunt large mammals, Cruxent joined a paleontological team excavating bones of mastodons and other extinct animals at the Muaco waterhole near the coast east of Coro. Much to his delight he found El Jobo points at Muaco, but also more recent artifacts; so it was clearly not possible to prove that El Jobo points had been used to kill mastodons at Muaco. Nevertheless, a fossil bone that evidently had been incised in a patterned way before permineralization was recovered at Muaco (Rouse and Cruxent l963, plate 4A), and burned bone from the site was dated 16,870 B. P. (Rouse and Cruxent l963: 35-36).

Encouraged by these finds, Cruxent searched for better sites where upwelling water might not have mixed materials. Two fossiliferous localities were found a few kilometres east, but one of these was a stream bed so it was concluded the El Jobo points and horse bones could have been introduced into the gravel deposit from different original contexts.

The more extensive Taima-Taima site is neither a waterhole nor a streambed, although artesian water does seep horizontally from the San Luis Mountains and flows through ancient beach sands beneath a bed of coquina containing Miocene fossils. Gradually the saturated sand caused the coquina to settle and crack. Water seeping through the cracks redeposited old beach sand onto the coquina pavement. The top of the pavement blocks were smoothed by the moving sand, but the water pressure was never strong enough to roll the pavement blocks. Animals, mostly mastodons, were attracted by water that pooled on the pavement in the grey sand. Early people were also attracted to the site, because Cruxent found El Jobo points in intimate association with mastodon bones, and one inside a pelvic cavity.

Several years after Cruxent's initial excavations, confirmatory excavations in l976 recovered a broken El Jobo point inside the pubic cavity of a partially disarticulated and butchered young mastodon whose bones had been cut, probably using a jasper flake found adjacent to the left ulna of the beast. The heavy bones had not been moved by water action, only by man. Preserved by constant wetness, a nearby concentration of twigs, evidently chewed by the mastodon and presumably part of its stomach contents, provided dates from several radiocarbon laboratories that indicated that the young mastodon had been slain and butchered about 13,000 B. P. (Bryan, et al. l978; Ochsenius and Gruhn l979; Gruhn and Bryan l984). The new dates confirmed most dates that had been obtained earlier on different materials from the grey sand. Eventually a soil developed on the grey sand, and horses and glyptodonts lived on this later surface, but evidently the mastodons had become extinct. Colluvium then accumulated and another paleosol developed on another surface. Mor colluvium was covered by an organic black muck, which yielded dates around l0,000 B. P. The entire depositional sequence was capped by sterile colluvium. No artifacts were found in any deposit above the grey sand, so the artifacts could not have intruded from above.

Despite claims to the contrary (e.g., Lynch l990), Taima-Taima has yielded definite artifacts in a well-dated, deeply stratified geological context which remained saturated with seeping water that moved sand and concentrated twigs into pockets, but was not strong enough to move or mix bones or stone artifacts. Reported in great detail, the data constitute the only solid evidence for a megamammal kill site anywhere in South America. Of course, if one accepts the model that North American big game hunters were the earliest South Americans, the evidence from Taima-Taima or any other site earlier than Clovis cannot be correct, and all such sites must be explained away, as Lynch (l990) has systematically attempted to do.

Significant differences between the archaeology of North and South America now become evident. North American archaeologists looking for cultural origins naturally look back to Beringia and Siberia; and it is quite easy to find similarities and relationships because people adapted to the Great Plains and the steppe tundra of Beringia and Siberia lived in ecosystems that required an emphasis on hunting as opposed to collecting. South American archaeologists naturally look back to Panama and Central America for cultural origins, but the dominant ecosystem in that region is tropical forest, and the only relationships that have been recognized are the widely scattered fishtail and Clovis-like fluted points. Archaeologists oriented toward North America thus conclude that fluted points must constitute the earliest relationships. Unlike fluted points, which have a widespread distribution because many cultural groups found them to be effective hunting weapons, El Jobo points have a very limited known distribution. Although willow-leaf-shaped points are known from Costa Rica, Nicaragua, and Mexico, they are not distinctive El Jobo points. Characteristic thick biconvex points like El Jobo have been collected from a site in north-western Argentina (Alberto Rex Gonzalez, personal communication, l970), and one was recovered in southern Chile at the contemporary 13,000 year old occupation site at Monte Verde to be summarized below. Future work, especially along the unknown eastern slopes of the Andes, may reveal a linkage between these widely separated occurrences; but it is also possible that the nearly cylindrical shape of El Jobo points was transferred from similarly shaped ground bone or wooden points in several parts of South America. A major difference between North and South America is the greater diversity in lithic assemblages which developed locally in South America as part of distinctive cultural adaptations to different ecosystems containing a wide and diverse range of natural resources. Technological convergence, including bifacial points with similar shapes, may be expected to occur as parallel adaptations to similar environmental and resource bases (Bryan 1973; Richardson l978).

In central Colombia, almost l000 km Southwest of Taima-Taima on the high Sabana de Bogotá, excavations in the El Abra rock shelters yielded a simple core and minimally retouched flake industry dated as early as 12,400 B. P. (Correal l986). The Abriense industry was also found at the open Tibitó site in association with remains of mastodons and horses but mainly deer in a context dated 11,740 B. P. (Correal l981). Tibitó may have been a processing station and not a kill site; but nevertheless the absence of any evidence of bifacial points, even broken, seems significant. These people probably killed their game with wooden spears, as did the occupants of Monte Verde, Chile, and as still do many lowland South American forest Indians. It is to be expected that Abriense people sometimes obtained bifacial points as trade items, but evidently these hunter/gatherers did not regularly use bifacially flaked stone projectile points on the Sabana de Bogotá, even into ceramic times.

The story near the equator in highland Ecuador is different. The region around Ilaló east of Quito contains abundant quantities of excellent obsidian, which attracted people at least by ll,000 B. P. Excavations at El Inga yielded several types of bifacial points, including fishtail fluted and unfluted forms; however, disappointingly late radiocarbon dates between 4,000 and 9,000 B. P. were obtained (Mayer-Oakes l986). Excavations at the nearby San José site by the American archaeologist William Mayer-Oakes yielded a similar industry, including blades, burins, and many carefully retouched unifacially flaked tools, but no evidence of bifacial flaking at all (William Mayer-Oakes, personal communication l990). A stratigraphically consistent series of obsidian hydration dates between 10,000 and 11,000 B. P. suggests that the earliest known occupants of the Ilaló region innovated an advanced flaked stone industry that lacked bifacial flaking only after l0,000 years ago. The idea of bifacial flaking was probably adopted when the Magellanic fishtail form diffused in from the south. Somewhat later the concept of fluting was accepted from the north and applied to fishtail points.

But Magellanic fishtail points are not ubiquitous farther south in the Andes, and they do not form a convenient horizon marker. In fact, they have a very spotty known distribution with gaps even at sites well dated to the time they should occur if the idea was passed from group to group. Rather, it seems that several projectile point styles were experimented with by different groups between 13,000 and 10,000 years ago, while other groups like those living on the Sabana de Bogotá never adopted the idea of bifacially flaked points at all, even though they hunted large mammals in a grassland environment. Unlike North America, where popular point styles became widely distributed, most South Americans continued to use functional technology that worked for them.

For instance, the earliest occupants of Pachamachay, a small rock shelter located at 4000 m on the high puna of central Peru, used a squat triangular point type to hunt vicuñas (Rick l980:149, fig. 7.1). By 9000 B. P. these people, who also hunted smaller game and gathered tubers and fruits, had switched to a peculiar willow leaf-shaped form with projecting bilateral spurs near the base, and stayed with that form for thousands of years. Evidently, these hunters successfully managed the vicuña herds to maintain their population, but began to herd the animals only about 4,000 years ago. The triangular points came from deposits dated 11,800 + 930 B. P., a date which Rick (l980:65) at first accepted, but later (Rick l988) cast doubt upon because of the large statistical error and the 2,500 year gap in time before the next occupation.

The story is quite different only l00 km north near Lake Lauricocha, the headwaters of the Rio Marañon. The first people who occupied Lauricocha Cave about 9500 years ago, immediately after glacial ice melted from the region, used abundant flakes, many with unifacial marginal retouch, as scrapers and points for hunting the Andean deer. Bone and antler points, identified by the excavator as daggers or awls, were more common artifacts (Cardich l978: 298). Bifacially retouched triangular and simple willow-leaf-shaped projectile points were used between 8,000 and 5,000 B. P., when people hunted more camelids than deer. The evidence from the high puna of central Peru indicates that neighbouring groups of people used different projectile point styles at the same time.

The possibility that bone points evolved into stone points at Lauricocha should be considered as an hypothesis which would explain local independent developments of bifacially flaked stone projectile points. Comparison of the early flaked triangular points at Pachamachay and Lauricocha with two ground bone triangular points recovered from beneath a massive rockfall in Pikimachay, a large rock shelter overlooking the Ayacucho Valley in southern Peru, suggests that a similar point form could easily have been transferred to a harder material when knappers familiar with simple marginal retouch experimented with retouching all edges on both sides to create a point that would penetrate thick hides more effectively than bone or antler points.

Pikimachay has become one of the more controversial sites in South America because evidence suggests that the shelter was occupied as early as 21,000 years ago.

However, as the flaked stones from the early deposits are of the same volcanic tuff as the cave wall, it is conceivable that roof spalls could have flaked when they struck the cave floor. Bones of extinct animals, mainly sloth, also appear to have been cut and worked, but this could also be ascribed to falling rocks. Nevertheless, unworked exotic pebbles which must be manuports have also been recovered from the same deposits. Although the equivocal unifacially flaked stones are found throughout the lower deposits, there can be no question that the triangular bone points from higher in the pre-roof fall deposits, which have been dated 14,150 + 180 B. P. (MacNeish, et al: 309, fig 8-l; l98l),are artifacts because the grinding marks are readily visible with the naked eye. One would have to suppose that several definite flake tools, and the ground bone points, all part of the Ayacucho Complex, were all intruded through the sterile layer of roof fall, a possibility that seems remote by the Canadian project geologist, Nathaniel Rutter (personal communication, l99l). Unifacial flake points were also described as part of the Ayacucho Complex; but the earliest biface, a rather crude percussion flaked bipoint, was found above the roof fall. Three broken points identified as "fishtail" appeared sometime after 11,000 B.P. (MacNeish, et al. l980). The reported evidence from Pikimachay makes sense if it is assumed that the earliest general hunter/gatherers carried with them an unspecialized unifacial flaked stone technology that embodied the potential to innovate bifacial points either by local experimentation or by external stimulus.

Although the earliest dated sites in western South America have been found on the puna where camelids abound and tubers can be collected, several localities have been found on or near the Pacific Coast that are more significant because of their artefacts content than their age. Shellfish, seabirds, fish, sea and land mammals, components of productive ecosystems especially at the mouths of rivers along the arid Pacific Coast, were exploited at least as early as l0,500 B. P. Significantly, the technology associated with these early coastal sites does not contain bifacially flaked artifacts, which would be expected if their immediate ancestors had been specialized big game hunters who occupied the high Andes.

On the arid Santa Elena Peninsula in Ecuador, Stothert (l985) reports on the Las Vegas culture, which is fully developed by l0,000 B. P., and earlier dates indicate its roots extend back another millennium. Shell, charcoal and human bone yielded stratigraphically consistent dates between 6,600 and l0,840 B. P. Artifacts include a unifacial flake industry lacking any formal types, although hammerstones, edge ground pebbles, and two ground stone axes were found. Bone points and a spatula may have been used for making nets or textiles. Scoops, dishes, and other containers were shaped of shell. The abundant industry of utilized but minimally retouched flakes and chunks suggests the manufacture of tools and equipment from wood, bamboo, reeds, and bark. Las Vegas is interpreted as an early tropical forest tradition which included some horticulture; and gave rise to the early ceramic Valdivia culture with intensive agriculture, an expanded fishing technology, and more developed ceremonialism after 5300 B. P.

The Talara tarseeps in extreme north-western Peru is a well-known Late Pleistocene paleontological locality. Nearby on the same marine platform, clusters of unifacial choppers, "horse-hoof" scrapers, denticulates, and utilized flakes were mapped and analyzed (Richardson l978). Associated shells of large mollusks, Anadara, evidently carried up from distant mangrove swamps, yielded dates of 11,200 and 8125 B. P. Probably the unifacial artifacts were used to work wood, bone, and fibres. With the addition of ground stone axes, mortars, and bowls, a similar industry continued to be used at the mouth of the nearby Siches River until at least 5500 B. P.

People adapted to the semiarid coast of the Santa Elena Peninsula and northern Peru evidently never found a need for bifacial points, although their absence seems strange if they had recently been big game hunters in the Andes. However, less than 500 km farther south, in the region around the Moche Valley, mastodons evidently were killed by hunters with distinctive large tanged Paiján points, which are known only from this limited coastal region, although a few similar forms were found at El Inga. In a sealed context in the small Quirihuac rockshelter, ten broken Paiján points and thousands of flakes were recovered. Four dates on wood and charcoal ranged from 12,795 to 8,645 B. P., while human bones dated 9,930 and 9,020 B. P. (Ossa l978). The open site of La Cumbre yielded Paiján points in association with mastodon bones that dated l2,360 and l0,535 B. P. As the dates were so variable, they were averaged to yield l0,796 B. P., as compared with an average of l0,650 for Quirihuac. It is believed that the Paiján Complex existed between 11,000 and l0,000 years ago, when the people gradually abandoned Paiján points but continued to emphasize the maritime economy that they probably pursued during a seasonal round (Richardson l989).

In southern Peru, near Ilo, marine resources were being utilized intensively by l0,500 B. P. A ring-shaped shellmound yielded a bone harpoon, bone and shell fishhooks, and modified shells, along with a unifacial flake industry, but no bifaces, in deposits dated between l0,570 and 7670 B. P. (Richardson l989). Mollusks, near shore fish, sea mammals, and bird remains were identified; but terrestrial mammals were absent.

Most likely, the reason a fully developed maritime adaptation was present on the Pacific Coast of South America by l0,500 B. P. and in California by l0,000 B. P., but not on the Atlantic Coast until after 8,000 B. P. is because sections of the Pacific Coast are tectonically rising, whereas the Atlantic Coast is stable. Thus, Las Vegas, Siches, and the Ring site all represent well established maritime adaptations which, like their contemporaries in southern California, happen to be the earliest locally preserved sites representative of more ancient coastal adaptations, remnants of which should be found by underwater archaeologists working on the continental shelves of all coasts. This interpretation implies that inland sites in the western cordilleras were first occupied by people who gradually expanded inland from the coasts as they adapted their economies to the utilization of terrestrial animals and plants. At first these explorers simply incorporated the resources of the adjacent interior regions in their annual round; but eventually some people, as at Pachamachay, developed a technology capable of adapting to year-round terrestrial ecosystems (Dillehay l989a). Two sites in central Chile have yielded artifacts with extinct animals. Quereo, located on a bluff now overlooking the Pacific, yielded simple unifacial flaked tools associated with mastodon, horse, extinct camelid, deer, and sea mammal bones in addition to marine shells in a context dated 11,500 years ago (Dillehay l989b). Evidently these coastal people were experimenting with interior ecosystems. Tagua-Tagua, farther inland in the central valley south of Santiago, is situated on a lake shore which attracted game (mastodons, horses, camelids, and aquatic birds) and hunters who left flakes, cores, hammerstones, and some bone tools in a stratum dated between 11,430 and 11,000 B. P. (Montané l968).

At Monte Verde, about 900 km farther south in the subantarctic rain forest and 15 km inland from the northernmost fjord, a boggy locality has revealed superbly preserved perishable artifacts in a context dated around 13,000 B. P. (Dillehay l989a, l986). The earliest known architecture in the Americas, at least ten semi-rectangular hut foundations made of roughly modified wooden logs held in place by wooden stakes; wooden mortars containing well-preserved seeds, fruits, and stalks of edible plants were directly associated with grinding stones. Wooden artifacts and a few unifacial flaked stone tools were found near small clay-lined hearths within the structures, and larger hearths were located outside the entranceways of these houses aligned along Chinchihuapi Creek. Monte Verde was a planned settlement with different activity areas for food preparation, tool production, and evidently even medical treatment. As plant remains were found which ripen in all seasons, it was concluded that the settlement was permanent. The presence of floral remains native to the ocean shores, the high mountains, and even Patagonia, indicate relations with other ecosystems, probably including trade. The presence of two large bifaces, and an El Jobo-like point of exotic materials probably were obtained through trade connections. Otherwise the lithic industry is composed of simple flakes and stones with naturally sharp edges, some of which were hafted onto wooden handles. The working edges of these carefully selected stones show clear traces of use (Dillehay and Collins l986). Grooved bolas stones and a wooden spear with a fire hardened tip were probably used to hunt extinct camelids and smaller game. Mastodon bones may have been collected as useful objects, although possibly people killed mired animals.

The final report on the artifacts from Monte Verde will give a clearer picture of the total material culture used by early hunter/gatherers. Test excavation across the creek in older sediments yielded several fractured stones, LL of which show clear percussion scars and/or use wear on their edges, in direct association with three hearth-like features that yielded dates of about 33,000 B. P. (Dillehay and Collins l988).

Many sites dated after 8500 B. P. have been excavated on the Chilean coast. These people collected shellfish, fished, and hunted sea mammals and birds, sometimes with bifacial willow-leaf-shaped and stemmed points. Some of these groups, especially in the arid north, moved upriver seasonally to hunt terrestrial animals and collect plants for food and medicine; however, most groups occupying the formerly glaciated south coast moved seasonally to various coastal localities. It has always been assumed that these Archaic maritime cultures had developed from earlier Paleo-Indian hunting cultures that used bifacially flaked stone projectile points, although the early evidence from Las Vegas, Talara, and the Ring site suggests that a general hunting/gathering economy had been followed in coastal zones before the innovation of bifacial points. The evidence from Monte Verde, Tagua-Tagua, Quereo, and Paiján suggests that early people living on or near the coast utilized all sizes of readily procurable game animals whether or not they used bifacial points. With this interpretation, the North American terms Paleo-Indian and Archaic, as well as their implied sequence, are clearly inapplicable in South America.

East of the Andes, people are known to have been on Tierra del Fuego, at the southern extremity of the Americas using a non-diagnostic unifacial flaked stone assemblage at least by 10,400 B. P. (Orquera l987). The earliest documented evidence of occupation of Patagonia is at Los Toldos, Cave 3 in interior southern Argentina, were the lowest cultural layer, Level LL, was dated 12,600 + 600 B. P. (Cardich l978; Orquera l987). Scrapers, knives, projectile points, and utilized flakes, all unifacially retouched, were found with bones of guanacos, extinct camelids, and horses. About 11,000 B. P. the shelter was reoccupied by the Toldense people, who used similar unifacial tools, with the addition of sub-triangular bifacial points and knives, as well as bone awls and spatulas, all associated with horse and guanaco bones. Toldense people left the region about 8750 B. P.; but people emphasizing unifacially retouched blades, scrapers, knives, and denticulates occupied the cave after 7260 B. P. Evidently these people used bolas rather than bifacial points to hunt guanacos.

Fell's Cave, in Chile north of the Straits of Magellan, is the type site for Magellanic fishtail points, two of which have long basal thinning scars. End and side scrapers were also recovered with broken and burned horse bones and many butchered guanaco bones in layers dated between 11,000 and l0,000 B. P. The subsequent occupation, dated between 9l00 and 8l00 B. P., reportedly lacked bifacial points. Bone points may have been used to hunt guanacos. Short triangular bifacial points and bolas stones were found in the third occupation dated between 8180 and 6560 B. P. The triangular points and overlapping dates suggest some relationship with Toldense (Orquera l987), although the lack of fishtail points in Argentine Patagonian sites of similar age remains a puzzle. Unlike North America, it seems that standardized projectile point styles were not widely adopted by contemporary groups. However, far to the north in central Buenos Aires Province, fishtail points were excavated from two nearby sites in contexts dated between l0,800 and l0,600 B. P. (Flegenheimer l987). The only identified bone was an extinct armadillo scute. Another locality in the region is apparently a workshop where fishtail points were shaped, one by grinding rather than flaking. Given the absence of kill sites, it is difficult to make a case for an early specialized hunting horizon simply from the known distribution of fishtail points in the southern cone (Orquera l987: 354).

Extinct glyptodon and guanaco bones were found with many exotic quartz flakes not far away at La Moderna. A bone collagen date of 6550 B. P. on the glyptodon has been questioned because it implies late persistence of extinct fauna (Orquera l987). However, Arroyo Seco in the southern part of Buenos Aires Province, has produced even greater surprises concerning extinct fauna. Bifacial triangular points occur with unifacial artifacts in an undated component. Below that stratum, a few tools with only unifacial marginal retouch were found associated with guanaco, deer, horse, and giant ground sloth bones. Beneath this occupation level and with no evidence of intrusion through the zone yielding the megafauna, were found human burials with red ochre, perforated shell and teeth beads, and a glyptodon scute. The suggestion that people maintained a graveyard while giant ground sloths and giant armadillo-like glyptodons were still living in the area appears to be confirmed by almost identical radiocarbon dates of about 8500 B. P. on sloth and human bones (Fidalgo, et al. l986). Certainly there is no evidence to support a model that big game hunters with a specialized technology had much effect on the Pampan fauna.

In southern Brazil, fishtail points are generally associated with more recent sites, although some stemmed points may be as early as 7,000 to 8,000 B. P. (Schmitz l987: 90). In fact, the only reported evidence for big game hunting in the savannah/parkland areas of Rio Grande do Sul are modified ground sloth bones, one of which yielded a date of 12,770 + 220 B. P. (Bombin and Bryan l978; Schmitz l987, fig. 14). Later people occupying this savannah/parkland region along the Rio Uruguai hunted deer and collected pitted fruits between 11,000 and 8,500 B. P. These hunter/gatherers used bifacially flaked stemmed points, bifacial knives, and a variety or flake scrapers, pounding stones, and anvil stones (Schmitz l987).

Along the Altoparaná in northeaster Argentina, eastern Paraguay, and south-western Brazil, a distinctive percussion flaked bifacial industry appeared sometime after 8,000 B. P. and lasted into protohistoric times. The large heavy bifaces, often looking like Old World handaxes and cleavers, led to the hypothesis that this industry was early and ultimately derived by diffusion from the Old World; however, apparently it was simply a local adaptation to a dense tropical forest ecosystem by hunters and collectors. Artifacts recovered include picks, scrapers, knives, heavy choppers and cleavers, but no bifacial projectile points.

Farther Northeast in the forested interior of the State of Sao Paulo, at the Alice Boër site near the city of Rio Claro, a unifacial core scraper and an end/side scraper

were recovered with several flakes on the surface of an alluvial gravel beneath a thick sterile alluvium. Unifacial artifacts and flakes were found throughout the clayey sand above the sterile alluvium. Tanged bifacial points appeared near the middle of the stratum and continued to the top. A small charcoal sample from the middle of the stratum yielded a date of 14,200 + 1150 B. P., and a burned chert flake from above this sample yielded a thermoluminescence date of 10,970 B. P., below charcoal dates which clustered around 6,000 B. P. (Bryan and Beltrão l978; Beltrão, et al. l986; Hurt l986). Evidence from Alice Boër suggesting that bifacial points were innovated as early as 14,000 years ago, earlier than anywhere else in the New World, must be confirmed at other sites; however, the presence of stratigraphically early unifacial artifacts seems to confirm the early dates in Rio Grande do Sul.

Farther Southwest in Sao Paulo State, but still on the Planalto, a series of occupation settlements, including open sites and rockshelters have been dated between 9,800 and 10,500 B. P. (Collet l985). Shells of giant land snails are the most abundant animal remains, although bones of tapir, deer, peccary, monkeys, small rodents, turtles and fish indicate a diversified hunting and collecting economy. Amorphous unifacial flake tools and pebble choppers were found, but bifacial points were absent. Ground mammal and bird bone points were used as projectile points. Flexed burials, one of which was dated 98l0 +_ 150 B. P., were placed in the occupation layers. The fact that these hunter/gatherers did not use bifacial points while their contemporaries to the Northwest and south used finely retouched tanged and stemmed bifacial points indicates that bifacial point styles were not widely adopted. Possibly these people seasonally climbed onto the Planalto from the coast where bifacial points were always very rare.

Shell mounds (sambaquis) were once common the coast from Rio Grande do Sul to Espirito Santo State north of Rio de Janeiro, although very few of them were excavated carefully before they were destroyed for construction materials. The largest sambaquis, more than 20 m high, were in the States of Santa Catarina, Paraná, and Sao Paulo, where there are many headlands, bays, inlets and estuaries which protected shellfish beds as well as the shellmounds from ocean swells. Because the Atlantic coast is quite stable, the gradually rising sea level reached its maximum after about 8,000 BP, so the submerged bases of the oldest known sambaquis, such as Maratuá near Santos, are dated after 7800 B. P. (Laming-Emperaire l968: 93-94). Most sambaquis date after 6,000 BP and had been abandoned before the introduction of ceramics in late prehistoric times. Flaked stone is usually rare; amorphous quartz flakes were used for cutting and scraping. Large percussion flaked bifaces were sometimes shaped as preforms for ground axes (Bryan l978). Pecked axes, chisels, mortars, netsinkers, pounders and hammerstones, often with pecked depressions for cracking palm nuts, also occur. Zooliths, beautifully polished stone representations of animals were reportedly associated with burials, but none have been properly excavated. Bone was used for projectile points, fish hooks, fish gorges, spatulas, beads, and pendants. Whalebone was made into braziers, containers, and other tools, including clubs and spindle whorls. Mammal and shark teeth were perforated for suspension; and shell was used for ornaments, knives, scrapers, and containers.

Because of the preponderance of shells relative to artifacts, the impression is gained that the occupants mainly gathered shellfish; however, the abundance of fish, and sea and land mammal bones in many sambaquis indicates that the people were very successful general hunters and gatherers. Nevertheless, huge sections of shellmounds often yield very few artifacts, an unattractive situation for archaeologists, so sambaquis were often ignored. But shellmounds are very significant because if it were not for preservation of the relatively abundant bone and shell artifacts, the few amorphous flakes and occasional hammerstone with pecked depressions (quebra cocos), coastal occupation sites may have gone unrecognized; and we would have had a very biased view of the prehistory of the productive Atlantic coastal zone, which was well populated in early historic times.

Future search by underwater archaeologists in protected and resource rich bays, lagoons, and estuaries should reveal older submerged sambaquis, possibly some containing waterlogged perishable materials.

Limestone caves and rockshelters have been excavated in the forested interior of the State of Minas Gerais since early in the last century. The original purpose was to procure bones of the Pleistocene fauna; however, human bones and artifacts were also recovered. The Danish palaeontologist, Wilhelm Lund, excavated many caves near Lagoa Santa, north of the city of Belo Horizonte. He concluded that the majority of human remains had been deposited long after the extinct ground sloths, mastodons, and other animals, except at one site where he found fossilized human bones in apparent association with extinct mammal bones. The human skulls from Sumidouro Cave, later studied by Pöch (l938), included some with heavy browridges, a feature that is absent on the later skeletons ("Lagoa Santa man") excavated by Lund and others. About the time of Pöch's study, other less well controlled excavations were carried out in Lund's and other caves in the area. The animal and human bones from these excavations were stored in crates in the State University in Belo Horizonte. Examination of the contents of the crates in l970 revealed evidence of human workmanship (chopping) on a few fossil bones (Prous l986), but also a human calotte with features similar to those included in Pöch's study except that the browrideges were even more pronounced (Bryan l978, figs. 7-12; Beattie and Bryan 1984). The calotte was misplaced before it could be studied properly, although the similar state of permineralization and the morphological features suggest that it came from Sumidouro Cave. The calotte may or may not date to the Pleistocene. Populations with heavy browridges are known elsewhere in the Americas from Holocene contexts. The calotte is significant not so much because it may be old but because it suggests that a transitional form of early Homo sapiens was present in America as well as in east Asia. In fact, the closest morphological similarities are with transitional Homo Sapiens in China and Australia.

Excavations by the French archaeologist, Annette Laming-Emperaire, in Lapa Vermelha IV, a large rock shelter in the Lagoa Santa region, finally confirmed the association of extinct fauna with human remains. Scattered bones of a giant ground sloth were recovered from a fissure at the rear of the cave, and human bones of the Lagoa Santa type without heavy browridges were found below in a level dated LL,600 and l0,200 B. P. (Laming-Emperaire, et al. l975; Prous l986a; Prous l986b). Minimally retouched quartz flakes and cores were excavated from a statum dated 22,400 in the main part of the cave, while a unifacially retouched side scraper came from just above charcoal dated >25,000 B. P. Excavations farther north in Minas Gerais yielded several quartz crystal flakes and red ochre associated with an extensive hearth dated 11,960 + 250 B. P. in the Gran Abrigo de Santana do Riacho (Prous l986a; l986b).

Many other caves and rockshelters have been excavated in Minas Gerais, Goias, and Bahia, although only preliminary and summary reports have been published. Some of these caves have yielded worked fossil bones of mastodon, sloth, and horse; but the lack of collagen in the bone has precluded radiocarbon dating. Charcoal dates on later levels are available back to about l0,000 B. P. The flaked stone industry from these caves is always unifacial. In fact, except for the occasional exotic trade item, the flaked stone technology remained unifacial until ceramics were adopted in the late First Millennium A. D. In most areas, shaped flaked stone artifacts are rare and late; however, in southern Goias a distinctive unifacially shaped limace (lesma in Portuguese) appeared between about 11,000 and 9,000 B. P. associated with unifacial choppers, knives, scrapers, and perforators. Heavy wear on one edge near one end suggests that the lesmas were used as scrapers (Schmitz, et al. l976; l989). Small and medium-sized mammals, reptiles, and birds, as well as fish and freshwater mollusks

were hunted, presumably with wooden spears, or gathered along with palm nuts, that were broken open with hammerstones containing pecked depressions on the sides. In the semiarid region of southern Piauí, several rock shelters at the base of a metamorphosed sandstone escarpment have been excavated near Sao Raimundo Nonato by a Franco-Brazilian team. The original intent was to date the rock art on the rear walls, but extensive and deeply buried evidences of occupation altered the prime objective, and there is now a research institute and a museum which is attracting tourists to the town (Guidon l986; Delibrias and Guidon l986; Delibrias, et al. l989; Parenti, et al. 1990). The Toca do Sitio do Meio was extensively tested in l978 and l980, but roof fall, consisting of large boulders, has prevented excavation of the lower layers. Nevertheless, pebble core choppers and unifacial flake tools were found associated with hearths dated 14,300, 13,900, 12,440, and 12,200 B. P.

Between l978 and l988, major excavations were carried out at the nearby Toca do Boqueirão da Pedra Furada because the roof of this large rock shelter collapsed only in front, creating a lip mound or rampart which would have screened occupants from view by anyone in the valley 20 m below. An alluvial fan, composed of quartz pebbles eroded from a formation high on the cliff face, has built up at one corner of the shelter. This fan contained a ready source of flakeable pebbles, although exotic quartzite pebbles were also found on several occupation floors, and chert was added in the upper layers. Except for one exotic bifacial point in an upper layer, the entire flaked stone industry is unifacial. Flake scrapers, notched flakes, pointed pebbles, pebble choppers, and hammerstones were identified, in addition to many cores and unretouched flakes, which are workshop detritus. A few artifacts examined under a scanning electron microscope revealed evidence of use striations, so the site is much more than simply a quarry/workshop where people also drew pictures on the walls.

The original objective of dating the rock art was successful. A sandstone wall spall bearing traces of red paint was found directly associated with a hearth dated 17,000 + 400 B. P. Extensive beds of charcoal, similar to hearths found elsewhere in Brazilian caves where people evidently dragged in limbs and logs to keep fires burning overnight, were found from near the surface to almost five m deep. The earliest people broke up some boulders with fire and rearranged them to make level areas for rock-lined hearths. Charcoal from this early horizon has been dated 41,000, 42,400 and  47,000 B. P. More than a dozen stratigraphically consistent radiocarbon dates on hearths built on later occupation floors have been dated between 32,160 and 6,100 B. P. The problem of the lack of bone preservation in the sandstone shelters has been overcome by excavations in nearby limestone caves, where bones of many Pleistocene animals have been recovered in addition to the fragment of a highly permineralized and thick-walled human calotte (Guérin l99l), study of which may help confirm the calotte from Lagoa Santa.

Announcement of these dates in preliminary reports has created consternation among North American archaeologists who have accepted the model that the Clovis people were the earliest Americans. None of these sceptic have visited the site, but they feel secure in their conviction that the artifacts must be naturefacts, the hearths are really from forest fires (e.g., Lynch l990), and even that the excavators are inadequately trained (Fagan l990b). Probably only a final report on the evidence from these sites will finally quell the controversy.

In the end these and all other sites dated as early as or earlier than Clovis will force rejection of the currently popular "Clovis-first" model and acceptance of an alternative explanatory model that does not have to explain away much of the actual reported archaeological data found throughout South, Central, and North America. Such a model would allow east Asians with a general hunting/fishing/gathering economy and a simple unifacial flaked stone technology gradually to expand their territory around the Northwest Pacific onto the unglaciated Bering Land Bridge, and then down the Northwest Coast of North America before it finally became glaciated. People with a maritime orientation would have stayed along the coast, although some groups would have budded off and moved up unglaciated river valleys, which would also provide productive ecosystems for general hunters and gatherers. Eventually, in a few open grassland areas which supported herds of herbivores having predictable habits but few edible vegetal foods, some people moving inland from the coasts and rivers would experiment with more effective methods for obtaining game, including innovation of bifacially flaked stone projectile points by a process of transference from similarly shaped wood and bone points. Exactly when the long process of the peopling of the Americas began will be determined by future archaeologists who are not hampered by a model with untenable assumptions that unduly restricts free scientific thought and action.