Pergamon PII: S1040-6182(97)00062 Quaternary International, Vol. 49/50, pp. 177-190, 1998. ? 1998 INQUA/Elsevier Science Ltd All rights reserved. Printed in Great Britain. 1 *" 1040-6182/98 $19.00 HUMAN SETTLEMENT OF CENTRAL AMERICA AND NORTHERNMOST SOUTH AMERICA (14,000-8000 BP) Richard Cooke Smithsonian Tropical Research Institute, Apdo 2072, Balboa, Panama A human presence in Central America, Colombia and Venezuela (outside 'Amazonia') before the Late Glacial Stage (% 14,000-10,000 BP) requires substantiation. In Venezuela, hunter-foragers who used elongate, thick bifacial projectile points preyed upon megamammals (e.g. gomphotheres) before the onset of a (as 12,000-11,000 BP) warming stage, particularly visible in highland Colombia (Guantiva Interstadial). Some archaeologists argue for a distinct pre-12,000 BP population centered on the Sabana de Bogota (Colombia, %2500 m). Their 'Abriense' stone tool kit lacked bifacial reduction, and continued to be used well into the Holocene. In Central America, no human cultural remains yet pre-date the widespread, but poorly documented 'Palaeoindian' horizon. This coincides chronologically with the % 11,000-10,000 BP cold stage equivalent to the Younger Dryas (in Colombia, 'El Abra Stadial'). Makers of lanceolate ('Clovis') and stemmed ('Fish-Tail') fluted points, carefully trimmed scrapers, burins and perforators, 'Paleoindians' have also been identified in northwest South America. Palaeoenvironmental data suggest that they moved around in many different vegetation types. In general, these were unlike modern potential vegetation because temperatures and annual rainfall were still lower than those of today. Human material culture at the onset of the Holocene (10,500 + 700 BP depending on region) does not seem not to have changed as rapidly as climate and vegetation. But the disappearance of fluted projectile points and the Late Pleistocene megafauna may be causally related. Presently, it is difficult to distinguish widespread and short-lived diachronic changes in human culture, from culturally diverse populations living synchronously in different geographic areas. Some archaeologists argue that early Holocene forest expansion and foraging territory contraction were responsible for socioeconomic diversification and a gradual shift from mobile hunting and gathering to a more sylvan way of life. This emphasized the intensive collection and incipient domestication of forest products (e.g. palms and root crops), and the hunting of deer and rodents. ? 1998 INQUA/Elsevier Science Ltd. All rights reserved INTRODUCTION This paper addresses the human settlement of Central America, Colombia and Venezuela (outside 'Amazonia') between ca. 14,000 and 8000 BP (Figs 1 and 2). A survey of the most relevant archaeological sites and artifacts is prefaced by vignettes of environ- mental changes inferred from palaeoecological studies. etation zones had moved close to their modern posi- tions. All palaeoecological data from below 1000 m come from areas with strongly or moderately seasonal climates, where forests doristically similar to modern mesic formations quickly replaced other LGS vegeta- tion types (i.e. temperate oak parkland, thorn-scrub, and herbaceous and wooded savanna). No palaeoeco- logical data refer to areas where modern vegetation is lowland evergreen rainforest. ENVIRONMENTAL CHANGE: GENERAL OBSERVATIONS The study region experienced a generally drier and colder (or cooler) climate during the Upper Pleniglacial (21,000-14,000 BP) (35-50% modern annual precipi- tation; atmospheric temperatures 8 + 1?C below mod- ern values depending on altitude (Leyden et ai, 1993; Markgraf, 1989)). Climate change during the Late Glacial Stage (henceforth LGS) (ca. 14,000-10,000 BP) continued to affect the vertical and horizonal zonation of vegetation in complex ways. Particularly relevant to archaeology is a warm-cold oscillation, which in time and inferred degree of temperature change is approx- imately equivalent to the warm-wet B0lling-Aller0d and cold-dry Younger Dryas (YD) stages. The oscilla- tion attains a high resolution in Colombia and Costa Rica above ca. 2000 m. It is less visible at lower ele- vations. The onset of the Holocene 10,500 + 700 BP (seemingly rapid in spite of geographic dating incon- sistencies) witnessed generally wetter and warmer conditions, although temperatures initially remained cooler than today. By 8500 + 500 BP, altitudinal veg- ENVIRONMENTAL VIGNETTES The following vignettes refer to geographical zones in which the archaeological sites discussed in the text are situated. Rapidly accumulating evidence argues for the interhemispheric coupling of Pleistocene stadials and interstadials (Lowell et ai, 1995) and for a global YD (Denton and Hendy, 1994). Over 35 yr have passed since van der Hammen and Gonzalez (1960) proposed that LGS climate change in the Cordillera Oriental approximated the Blytt-Sernander sequence. (For sake of convenience, the Guantiva Interstadial (ca. 12,000- 11,000 BP) and El Abra Stadial (ca. 11,000-10,000 BP) of the Colombian Andes are employed henceforth as regional chronostages even though this is not formally correct for the South American lowlands or for Central America.) 177 178 R. Cooke ARCHAEOLOGICAL SITES ft FLUTED POINT FIND-SITES * CO/l/A/GLOC/IL/r/?S + ALTITUDE (m ABOVE SEA LEVEL) FIG. 1. Central America (Huehuetenango: see Hayden, 1980). The best evidence for Pleistocene-Holocene culture change comes from a high intermontane plain known as the Sabana de Bogota (2500-2600 m). Maximum depression of Pleistocene glaciers (3200 m) occurred ca. 21,000 BP, before a human presence can be demon- strated. LGS vegetation varied between a cold-dry extreme of herbaceous paramo to a warm-wet extreme of Andean forest (Alnus, Myrica, Symplocos, locally Quercus and Weinmannia), with intermediate semi-open subparamo dominated by Compositae. The last-named vegetation type probably existed near the El Abra and Tequendama rockshelters during the El Abra Stadial, when inferred local annual temperature was 4-6?C below modern values. Higher precipitation (i.e. during the Guantiva Interstadial (temperature 2?C lower than today) expanded marshes and small lakes in the basin of 'Lake Bogota', which drained 30,000 BP. During the Early Holocene Andean forest again invaded the sabana to ca. 3000 m (van der Hammen and Correal Urrego, 1992; van der Hammen and Hooghiemstra, 1995). Colombian Subandean forests (1200-2300 m) Calima valley palynology suggests a gradual, but oscillating, LGS/Early Holocene withdrawal of Quercus as Andean forest moved upslope. This landscape between 14C core-section dates of 9590 + 150 and 6680 + 230 BP was heavily forested, but the boundary between Andean and Sub-Andean forest remained close to Sauzalito and El Recreo (see below) until the latter date (Herrera et al., 1992). The Magdalena lowlands (Colombia) The middle Magdalena valley (altitude: 150-600 m, precipitation: 2500 mm/yr) was a focus for LGS and Early Holocene hunter-gatherers. All known sites ap- pear to post-date the Guantiva Interstadial, during which there is geomorphological evidence for a locally moist climate. During very cold Late Pleniglacial or LGS Ciega Stadial (ca. 13,000-12,000 BP) times, an altitudinally continuous belt of open vegetation probably connected the lower valley with the eastern Cordillera (thereby facilitating hunter-gatherer mobility) (van der Hammen and Hooghiemstra, in press). During the less cold El Abra Stadial, lowland vegetation perhaps remained xeric and relatively open (Lopez Castano, 1993). Lake Valencia, Venezuela The coastal plateau where this lake is located (alti- tude: 1000 m, precipitation: 900 mm/yr) was adjacent Human Settlement of Central America and Northernmost South America 179 Tequendama 0 Km 150 FIG. 2. Colombia and northern Venezuela. to herbaceous grassland and/or dry thorn scrub 13,000-10,900/10,500 BP. Mesic forest, which en- croached upon the watershed before the latter dates, became dominant 8300-7500 BP (Bradbury et al, 1981; Leyden, 1985). The more arid Venezuelan coastal margin ? where hunter-gatherers killed or scavenged extinct megammammals at springs ? probably experi- enced desert-like conditions during much of the LGS. The Central American highlands The Cordillera de Talamanca (western Panama/ eastern Costa Rica) and the Sierras Madre and de las Minas (Guatemala) rise to over 3000 m. Both ranges were glaciated during the Late Pleistocene, when tree line dropped 600-800 m (Horn, 1990; Hastenrath fide Schubert, 1988). Pollen evidence from the La Chonta bog (Talamanca, altitude: 2310 m) suggests a YD- related cooling event 11,100-10,400 BP, with reduced rainfall and inferred annual temperatures 2-3?C lower than present values. Holocene conditions are evident by 9800 ? 120 BP (Islebe et al, 1995). The Central American lowlands Palaeoecological records from Lakes Quexil and Salpeten (Guatemala) and Lake La Yeguada (Panama) suggest that LGS vegetation change was more uni- directional than above 2000 m, although Horistic varia- bility during the LGS may have been climate-induced (Bush et al, 1992; Leyden et al, 1993). Even so, LGS hunter-gatherers in Central America would have en- countered vegetation taxonomically and areally differ- ent from today's potential formations, due to lowered 180 R. Cooke atmospheric temperatures (5+1.5?C), and reduced and/or more seasonal precipitation (Piperno (1995) re-interprets proxy indicators of LGS aridity at La Yeguada; cf. Bush et al. (1992)). In the Guatemalan Peten, oak parkland replaced Late Pleniglacial xeric scrubland (Brenner 1993; Leyden et al., 1993). In the Pacific foothills of Panama (altitude: 500-650 m), oak- Ilex-magnolia forests persisted until the onset of the Holocene, dated at La Yeguada to ca. 11,000 BP (Bush et al., 1992). Piperno et al. (1991a) inferred from the La Yeguada and El Valle cores (Bush and Colinvaux, 1990) that the seasonally arid Pacific lowlands of Panama ? and, by extrapolation, much of the Pacific slopes further north ? were covered by thornscrub and herbaceous savannas, in places expanded oceanwards by still lowered sea levels. This open landscape would have extended through lowland Darien into Colombia. Data from fossil Monte Oscuro lake, Panama (altitude: 20 m, precipitation: 2000 mm/yr), substantiates their hypothesis. Here, a Pleistocene palaeosol, which prob- ably formed under grassy thorn-scrub vegetation (with Curatella and Arundinella), is stratified beneath Early Holocene mesic forest (Piperno, 1995). no retouch, but are sometimes trimmed (usually on one edge) by direct percussion (Fig. 3b). Wood- working is assumed to have been a dominant activity. (2) More 'specialized' hunters with extensive foraging ranges, who made bifacial projectile points and cutting tools, as well as unifaces sometimes re- touched with controlled percussion flaking. Prey animals included extant taxa (i.e. deer) as well as extinct ones (i.e. gomphotheres, ground sloths, horses, and camelids). Late Pleniglacial antecedents? Correal Urrego (1993) reported flaked stone artifacts sealed under LGS volcanic ash and colluvium at Pubenza (Cundinamarca, Colombia), in the Mag- dalena river lowlands. These include an obsidian spent core and a chalcedony flake with a flat striking plat- form (Fig. 3a), associated with a dispersed wood char- coal date of 16,400 + 420 BP. Faunal remains include mastodont, as well as 'turtle, snails, crabs, mouse, and armadillo'. ARCHAEOLOGY More than one Pre-Palaeoindiari culture? In North America, discussions about initial human colonization still revolve around the 'Clovis' culture. This first stage of the 'Palaeoindian' period is custom- arily associated with hunting medium- and large-sized animals ? including extinct megaherbivores ? in open landscapes. 'Clovis' tool kits are widespread vir- tually throughout the unglaciated (and some of the deglaciated parts) of this continent between ca. 11,200 and 10,600 BP (Meltzer, 1995). A year after the first description of a 'Clovis' site, a different kind of stemmed ('Fish-Tail') fluted point was found in rock shelters in Chilean Patagonia along- side bones of extinct horses (Onohippidium \Parahip- parion\) and giant ground sloths (Mylodon). If two University of Pittsburgh dates are ignored as outliers (cf. Borrero, 1995), the 'Fish-Tail' fluted point tradition in southernmost South America lasted from ca. 11,170 to 9920 BP (Bird, 1988; Politis, 1991; Table 1). The view that a fluted point horizon derived from 'Clovis' is also the oldest human culture (Lynch, 1990), is not the majority opinion among archaeologists who work in South America (Bryan, 1991; Dillehay et al., 1992; Gruhn and Bryan, 1991). There is now a general consensus that at least two culturally and adaptively distinct ? and wholly or partially coeval ? 'popula- tions' or 'traditions' were present in this study area before the Palaeoindian horizon: (1) 'Generalist' hunter-foragers who moved around in small groups. Their stone tool kit lacks bifacial reduction, and consists of flakes and core tools made from chunks or water-worn nodules of siliceous materials. Working edges generally have Pre-Paleoindian predation on extinct fauna Colombia 14C dates are pending for possibly early LGS sites at Totumo (Cundinamarca) and Las Piletas (Santander), which contain megafaunal remains, flakes, and cores. At Totumo, a circular feature containing Haplomas- todon and sloth teeth from adults and juveniles recalls features from Tibito in the Sabana de Bogota (Correal Urrego, 1993). Tibito (altitude: 2600 m) is in the low dry montane forest zone of the Eastern Cordillera. In a 140 m2 decapage, Correal Urrego (1981) identified human activities in association with extinct proboscideans (Cuvieronius and Haplomastodon) and horse (Equus (Amerhippus), MNI = 8), and extant deer (Odocoileus) and fox (Cerdoycon). Stratum 3A ? a 0.2 m deep sandy loam ? contained three features, in which burnt and unburnt horse and gomphothere bones were mixed with stone artifacts and chunks of limestone. The pre- ponderance of teeth and ribs suggests in situ dismem- berment and joint-removal (the animals could have been scavenged, but site geology does not suggest ponding). The single 14C (bone) date of 11,740 ? 110 BP (Grn 9375) is concordant with the palynology. Stratum 3A's inferred oak-alder wooded parkland is associated with the Guantiva Interstadial at this alti- tude. Overlying Stratum 3's more herbaceous land- scape correlates with the El Abra Stadial. Correal Urrego attributes all 156 stone artifacts to a Hake-core tool tradition called 'Abriense' in Colombia (Correal Urrego et al, 1966-69), except for a 'finely retouched keeled' scraper, which he attributes to the partially contemporaneous 'Tequendamiense' industry 62% are 'used flakes' and 'core fragments'. 78% of the artifacts were found in Stratum 3A's circular features. Human Settlement of Central America and Northernmost South America 181 FIG. 3. Stone tools from Colombian sites (not to scale; size of each tool given below) (a) Pubenza, possibly Upper Pleniglacial (W: 35 mm; photo: R. Cooke, courtesy Institute Colombiano de Antropologia, (b) Totumo (Cundinamarca) (L: 50 mm; photo: R. Cooke, courtesy Institute Colombiano de Antropologia), (c) 'Hoe', Sauzalito (L: ca. 150 mm), (d) 'Hoe', El Recreo (L: ca. 145 mm; after Gnecco Valencia and Salgado Lopez, 1989: Fig. 3, 1 and 3), (e) biface, Tequendama II (L: 62 mm, photo: R. Cooke, courtesy Institute Colombiano de Antropologia), (f) re-used bifacial point fragment, Tequendama II (L: 41 mm; photo: R. Cooke, courtesy Institute Colombiano de Antropologia), (g) discoidal stone, San Isidro (Diam: ca. 95 mm; photo: R. Cooke, courtesy C. Gnecco Valencia), (h) smoothed axe, San Isidro (L: ca. 130 mm; photo: R. Cooke, courtesy C. Gnecco Valencia), (i) end-scraper, Tequendama II (L: 46 mm; photo: R. Cooke, courtesy Institute Colombiano de Antropologia), (j) lanceolate biface, San Isidro (L: ca. 120 mm; photo: R. Cooke, courtesy C. Gnecco Valencia), (k) unifacial blade tool, La Palestina (Magdalena valley) (L: 105 mm; photo: R. Cooke, courtesy C. Lopez Castaho), (1) unifacial scraper, Penones de Bogota (L: 92 mm). Northern Venezuela Correal Urrego did not find bifacial tools at Tibito. At Monte Verde (southern Chile), however, two elon- gate lanoceolate points were found in a 5-10 cm-thick layer stratified underneath peat dated to 11,760 + 470 and 10,860 +130 BP. Culturally modified animal bone from the same context ? including mastodont (Cuvieronius) and camelid (Paleolama) ? was directly dated to 12,000 + 250 and 11,990 ? 200 BP (Dillehay and Collins, 1991; Dillehay et al, 1992; Fig. 3). Long, thick, 'Joboid' lanceolate points are wide- spread in Venezuela (Fig. 5a; Ardila Calderon and Politis, 1989; Rouse and Cruxent, 1963). A 'Joboid' point mid-section was found within the obturator fora- men of a juvenile Haplomastodon at Taima-Taima (Venezuela), an artesian spring just inland from the present-day beach (Bryan et al, 1978). In the same stratum were glyptodontid, horse, mylontid sloth, ur- sid, and felid bones. Taima-Taima's excavation and conservation history (Gruhn and Bryan, 1984, p. 128; Cruxent, 1979, p. 77, note) have jeopardized objective interpretations (Morlan, 1988, pp. 35-36). The Monte Verde finds, however, reaffirm this site's importance as a pre-Palaeoindian locality. Well-dated Early Holo- cene lanceolate points in the Andes (Lynch, 1980, pp. 186-191) do not invalidate a pre-'Clovis' date for the Monte Verde and Venezuelan 'Joboid' points. Since the latter differ technologically and morphologically from Early Holocene lanceolates, it is reasonable to assume that they were designed for different tasks. The dis- patching of mortally wounded megamammals mired in muddy springs with lances tipped with long, thick, socketed bifaces, is not an intemperate inference. A Iff range of 13,520 and 12,430 BP (13 dates) for the convol- uted grey sand in which the gomphothere reposed at Taima-Taima (Bryan and Gruhn, 1979) suggests a pre- or early Guantiva Interstadial age. Palaeoenvironmen- tal data are, however, inconclusive (Gruhn and Bryan, 1984). Conditions as, or more arid than today's, would have forced prey animals to congregate at perennial springs (Oschenius, 1979; cf. C.V. Haynes, 1991; G. Haynes, 1991). Recent geomorphological research and thermoluminescence dating also suggests that 'Joboid' sites pre-date 12,000 BP at the Piedra de Chispa site (Oliver and Alexander, 1990). 182 R. Cooke LGS 'Abriense' occupation of Sabana de Bogota rockshelters The 'Abriense' tool tradition was first identified at 'El Abra' ? really three small shelters (2, 3, and 4) located in two parallel sandstone escarpments in the Sabana de Bogota (altitude: 2570 m). The small 'Abriense' sample from LGS deposits comprises rough flakes and angular fragments of brown and black chert considered 'non- local' in reports, but, apparently, available in river terraces near the site (Hurt et al., 1977, p. 14). Some flakes are pointed and blade-like. A few pieces with percussion retouch on one edge are described as scrapers, knives, and spokeshaves (Correal Urrego et al, 1966-69; Hurt et al, 1972, 1977). Charcoal samples were collected at El Abra from arbitrary levels, which were then related to natural stratigraphy (Correal Urrego et al, 1966-69: Fig. 5; Hurt et al., 1997, p. 4). Summary articles (e.g. Correal Urrego, 1986; Reichel-Dalmatoff, 1992 (pp. 44-45)) attribute the LGS 'Abriense' only to the 12,400 + 160 BP date. This is, in fact, the oldest of four dates obtained in most informative rockshelter (# 2), within or at the upper boundary of deposition unit (DU) C3, where the earliest formally described human occupation also occurs. The other dates are: 11,210 + 90, 10,720 + 400, and 8810 +420 BP. The charcoal for the 12,400 BP date was collected in arbit- rary level 7, along with nine stone artifacts. A date of 9420 + 110 BP was obtained in arbitrary level 8 under- neath it. It was discarded by Correal Urrego et al. (1966-69, p. 16), because they feared rootlet contami- nation. The date is not mentioned in Hurt et al. (1997), van der Hammen's (1978) profiles or summaries of Colombian 14C dates (e.g. Botiva Contreras, n.d.; Uribe, 1990). Pollen distribution inside the rockshelters is discon- tinuous. LGS and Early Holocene strata in shelter # 2 lacked pollen. A Guantiva Interstadial age and envi- ronment were attributed to DU C3 by stratigraphic cross-referencing with pollen cores taken in or near a gully in front of the shelter, and also from borings in the valley bottom (Schreve-Brinkman, 1978; van der Hammen, 1978: Fig. 27). The two dates accepted by the excavators for DU C3 ? 12,400 and 12,210 BP ? are consistent with this palaeoenvironmental reconstruc- tion. In DU C4, above C3, cross-referenced pollen data point to an El Abra Stadial sub-paramo vegetation. The 10,720 BP date collected at the junction of C3 and C4 is confirmatory of this age. DU D^ accumulated above C3 during the 10th millennium BP. In these earliest Holocene times, annual chert deposition in rockshelter # 2 was considerably greater, and occupa- tion either more regular or more intense than during the LGS (Hurt et al, 1972, 1977, p. 6). In sum, even though people appear to have occupied the El Abra shelters during the LGS, they went there sporadically or for very short periods of time. Incon- sistencies among the cultural, chronometric, and palynological data warn against emphasizing a 13th millennium BP origin for the 'Abriense' edge-trimmed tool 'tradition' at this particular locality. Remarkably consistent radiometric dates, greater density of occupation debris, and clearer cultural fea- tures were recorded at the nearby Tequendama II rock- shelter (altitude: 2570 m), where the 'Tequendamiense' lithic industry was first described (Correal Urrego and van der Hammen, 1977). There is no evidence for pre- LGS human activities. LGS cultural remains are con- centrated in DU5, a 30 cm-thick humic, and slightly sandy soil, which has two sub-units. Scant 'simple flakes and a few artifacts' were found in DU 5a, but have not been described (Correal Urrego and van der Hammen, 1977) (pp. 34, 167). Immediately overlying level DU 5b contains shelter exfoliation debris, ash from hearths, and diminuted bone. It is anchored by three dates: 10,920 + 260 BP (junction with underlying DU 5a), 10,730 + 105 BP, and 10,460 + 130 BP. Five charcoal samples from its boundary with DU 6, which represents the LGS/Holocene transition, overlap at la (range: 10,300-9890 BP). DU 5b is indisputably of cul- tural origin and of El Abra Stadial age. Correal Urrego and van der Hammen assigned DUs 5 and 6 to 'Occupation Zone F. The 'Abriense' tool class in this zone appears better made, and morpho- technologically more uniform, than the small LGS 'Abriense' sample at El Abra. It is very similar to the earliest lithic sample from Nemocon IV (Correal Urrego, 1979). Only three bifacial tools were found (Correal Urrego and van der Hammen, 1977, p. 68). One appears to be a blade fragment of a thin chal- cedony bifacial point, re-used on two edges (Fig. 3f). Its flaking pattern recalls that of an out-of-context point from Bahia Gloria (Fig. 5b: Correal Urrego and Pinto Nolla, 1983), and some of the thinner Central American 'Clovis'-like fluted points (e.g. Fig. 4c). The others are a thin quartzite foliate biface and a crude percussion- flaked waisted biface (Fig. 3e). A crude unifacial ovate scraper (limace), a large flake with semicircumferential retouch, and a stubby unifacial end-scraper (Fig. 3i) were also assigned to the ' Tequendamiense' industry. The latter recalls end-scrapers from the Los Tapiales 'Clovis'-like assemblage (Gruhn and Bryan, 1977). A similar tool was found at Tibito (Correal Urrego, 1981; photo 13). Four of these ' Tequendamiense' artifacts were found together ? as though they had 'been forgotten by their owners'. Hence, they should be coeval (Correal Urrego and van der Hammen, 1977, p. 155). The unification of DUs 5 and 6 for analytical purposes implies that they may be as late as 9500 BP. The presence of pressure- flaking is doubtful (contra Correal Urrego and van der Hammen, 1977, p. 167). Only the re-used bifacial point fragment and the end-scraper seem 'carefully made'. Several cooking, refuse disposal, and 'work' areas were identified at Tequendama II (van der Hammen 1977, pp. 156-160). In 'Occupation Zone F, mammal bones ? including a few simple tools ? were concen- trated in one refuse feature. No extinct taxa were re- ported. The following rank-order of abundance can be Human Settlement of Central America and Northernmost South America 183 FIG. 4. Stone tools from Central American sites, (a) Fluted preform, Finca Guardiria, Costa Rica (Snarskis, 1979: Fig. 4b), (b) lanceolate 'Clovis' point, Ladyville, Belize (Kelly, 1993: Fig. 5a), (c) 'Waisted Clovis' point, Lake Madden, Panama (Bird and Cooke, 1977; Fig. 3c), (d) complete Fish-Tail point, Lake Madden, Panama (Bird and Cooke, 1977: Fig. 4e), (e) fluted point base, Los Tapiales, Guatemala (Gruhn et al, 1977: Fig. 10a), (f) resharpened 'Fish Tail' fluted point, Lake Madden, Panama (Ranere and Cooke, 1995: Fig. 1.2d), (g) perforator on bifacial thinning flake, Los Tapiales, Guatemala (Gruhn et al, 1977: Fig. lie), (h) end-scraper with lateral spurs, La Mula-West, Panama (Ranere and Cooke, 1995: Fig. 1.1b), (i) bifacial thinning flake, Lake Madden, Panama (not to scale, max. width: 68 mm; photo M. Guerra), (j) unifacial tool, Los Tapiales, Guatemala (Gruhn and Bryan, 1977: Fig. 11a), (k, 1) burins, Los Tapiales (Gruhn and Bryan, 1977: Fig. lie and d). inferred from one analyzed sub-sample: [1] white- tailed deer (Odocoileus), [2] rabbit (Sylvilagus), [3] cot- ton-rat (Sigmodon), [4] wild guinea-pig (Cavia), [5] armadillo (Dasypus), [6] brocket deer (Mazama), [7 =] tayra (Eira) and kinkajou (Potos). (Skunk [Conepatus] and opossum [Didelphis] abundance is unclear) (Correal Urrego and van der Hammen, 1977, pp. 48-50). Some taxa (e.g. Potos, Eira) are unlikely to have occurred in herbaceous sub-paramo. If Andean forest descended 300-600 m during the El Abra Stadial (Hooghiemstra and van der Hammen, 1993, pp. 258-259), hunters would have travelled a considerable distance to obtain these animals. The Sueva rockshelter (altitude: 2690 m) also pro- vides evidence for a quite intensive, but apparently short-lived human occupation. A 0.2-0.5 m dark black layer (DU 3), with a single 10,090 + 90 BP date, con- tained > 5200 'Abriense' artifacts and > 300 animal bones of which 98% are white-tailed deer (Odocoileus). The only botanical information is a fragment of burnt forest bamboo (Chusquea). 46% of DU 3's lithics are described as waste flakes, 6% as scrapers, and 2% cores and core fragments. There is no evidence for bifacial flaking. A primary flexed adult female was found buried in DU 3, associated with an offering (?) of 14 stone flakes, a 'side-scraper', and ochre and hematite fragments. It is the only possibly LGS human burial from the study area. Underneath DU 3 at Sueva is a 0.1-0.6 m silty clay layer (DU 2). It has not been dated, but its pollen suggests a damp environment of alder forest intermixed with open herbaceous areas. It contained more than 1000 artifacts, and 170 animal bones (96% deer). Correal Urrego (1979) (p. 245) concludes that DU 2, which was in places sealed by sterile roof-fall, corresponds to the Guantiva Interstadial. However, the edaphological and cultural differences between this layer and DU 3 could also be due to varying occupation intensity. If the El Abra Stadial terminated in the northern Andes by 10,000 BP (van der Hammen and Hooghiemstra, in press), DU 2 and DU 3 may be transitional Pleistocene/Holocene, rather than LGS in age. 184 R. Cooke FIG. 5. Colombian and Venezuelan projectile points, (a) 'Joboid' point, Pedregal valley, Venezuela (Szabadick collection; photo: J. Oliver); (b) Cueva de los Murcielagos, Bahia Gloria (photo: J.B. Bird), (c) San Isidro, Cauca valley (photo: C. Gnecco Valencia), (d) San Isidro, Cauca valley (photo: C. Gnecco Vallencia), (e) El Cayude, Paraguana Peninsula, Venezuela (photo: J. Oliver), (f) Restrepo, Calima river, Colombia (L: 91 mm) (Reichel-Dalmatoff, 1986: Fig. 3), (g) Las Piletas, Santander (photo: R. Cooke, courtesy of G. Correal Urrego), (h) Estado de Falcon, Venezuela (photo: A. Jaymes), (i) Yondo, Magdalena river, Colombia (photo: R. Cooke, courtesy G. Correal Urrego). Fluted points: morphology, technology, and time I do not believe that 'fluting' was a 'minor technolo- gical tradition' (Gruhn and Bryan, 1977). It is a funda- mental aspect of composite projectile technology, which involves both perishable and non-perishable parts (Bird, 1969; Frison, 1989). Fluting was not in- vented in the Old World. Wherever it was first de- veloped and whichever direction its dispersal followed, allopatric convergent evolution in North and South America is unlikely. In this study area, only the Los Tapiales lanceolate basal fragment (Fig. 4e) was found in 14C-dated buried deposit. The complete point from Ladyville (Belize) (Fig. 4b (from Kelly, 1993: Fig. 5)) and one from Finca Guardiria (Ranere and Cooke, 1991: Fig. 6a) would probably satisfy Palaeoindian specialists' formal cri- teria for 'Clovis'. Some fluted points from El Cayude (Venezuela) (Fig. 5e; Oliver and Alexander, 1990; Ardila Calderon and Politis, 1989) probably belong in this group. Other complete points from (1) San Rafael (Guatemala), (2) Site G-164-A5 (Costa Rica) (Sheets, 1994), (3) Guanacaste (Costa Rica) (Bird and Cooke, 1978: Fig. lb), (4) Finca Guardiria, and (5) Lake Madden (Panama) (Fig. 4c), are sub-medially re- stricted, and have outflaring 'ears' at the base (often broken). Denominated 'waisted Clovis' (Ranere and Cooke, 1991), they recall regional Palaeoindian vari- ants in the US, e.g. Florida 'Suwannee' and 'Simpson' points (cf. Dunbar, 1991: Figs lc and 2b). Five stemmed fluted points from Lake Madden (Panama) (Fig. 4d and f; Bird and Cooke 1977, 1978), and one from Finca Guardiria (Snarskis, 1979: Figs 2f and 3a), have the base-blade junction intact. Their dimensions are similar to those of Ecuadoran and southern South American 'Fish-Tails', but they have straight rather than outflaring 'ears'. Stems are but a fraction wider. This homogeneity alludes to a func- tional relationship between stem and blade width, and shaft circumference, upon which projectile penetration and strength is predicated (Bird, 1969; Politis, 1991) (p. 295). On both lanceolate and stemmed fluted points, the edges of basal sections are abraded to prevent them from cutting into the hafting medium. 'Restrepo' fluted points have been found only in Colombia, one in an (undated) lowland cave (Cueva de los Murcielagos, Bahia Gloria (Correal Urrego and Pinto Nolla, 1983)). They have a narrow base with a straight bottom edge, an elongate triangular blade, and a right-angled stem-blade junction. One has Human Settlement of Central America and Northernmost South America 185 prominent serrations. Basal bilateral fluting channels sometimes extend to the middle of the blade (Fig. 5f and g; Ardila Calderon, 1991: Fig. 4(2-4); Reichel- Dalmatoff, 1965: Fig. 8). One broad-stemmed point from La Elvira (type 2A) is fluted (Gnecco Valencia, 1994). Fluted point morphology appears to correlate with two tool reduction processes. Lanceolate and, appar- ently, 'Restrepo' point manufacture, began with a bi- facially prepared blank, which was then thinned by percussion flaking. Thinning flakes sometimes continue from edge to edge; mistakes make elongate 'overshoot flakes' (Ranere, 1992: Fig. 2j; Ranere and Cooke, 1995: Fig. Lie). The cross-sections of mid- (unfluted) sections are bi-convex. Conversely, 'Fish-Tail' stemmed points, and the La Elvira stemmed example, are made from flake blanks that are not much thicker than the finished point. Thinning flakes 'tend to broaden at their distal end and overlap each other in the center of the point' (Ranere and Cooke, 1991, p. 239). Three lithic samples point to the mutual exclusivity of the two reduction processes. At Los Tapiales, where biface-dressing is judged to have been a primary activ- ity, lenticular platforms prepared by grinding or crush- ing occur, but 'lips' are rare. At La Mula West (Panama), perhaps an artifact refitting station, 67 biface fragments, scores of bifacial thinning flakes, and 'over- shoot' flakes, and six fluted point fragments bear wit- ness to the thick bifacial blank technology (Cooke and Ranere, 1992b; Ranere, 1992; Ranere and Cooke, 1995). On the other hand, abundant bifacial workshop debris from the Lake Madden (Westend) site ? close to the reported fluted point find-sites (Bird and Cooke, 1977) ? uniquely represents the large flake reduction tradi- tion. Thinning flakes have tiny, abraded and lipped platforms. Many are triangular (Fig. 4i). An alternative explanation for these differences is that they are related to raw material availability and workability, rather than a technology/time/culture in- terface. Fully 77% of Los Tapiales' artifacts are basalt; almost all La Mula West's are banded agate. These are perhaps inappropriate materials for taking off large thin flakes, unlike the fine Lake Madden chalcedonies or the La Elvira obsidian and glassy chert (Gnecco Valencia, 1994). Nevertheless, closer attention to the relationship between technology and morphology is encouraged in order to determine whether fluted point distributions exhibit an areal diachronic evolution (i.e. 'Clovis' and 'Restrepo' before 'Fish-Tail' and La Elvira' type 2A) or a regionally coeval distribution (i.e. different populations with contrasting tool manu- facturing processes) (Dillehay et al., 1992; Gnecco Valencia, 1994). Gruhn and Bryan (1977, p. 244) believe that 10,700 BP is a 'reasonable approximate estimate' for 'Clovis' Los Tapiales and nearby La Piedra del Coyote (Guatemala). La Elvira (where only one point, in a sample of more than 100, was fluted) is assumed to be coeval with or slightly antecedent to nearby San Isidro (ca. 10,000-9500 BP) (Gnecco Valencia, 1994). Fluting is absent at San Isidro, and also at San Juan de Bedout, La Palestina, and Nare in the lower Magdalena valley (la range, four dates: 10,490-10,150 BP) (Lopez Castano, 1995a, b). These tenuous contextual data ? in addition to radiometrically dated associations from outside this study area ? suggest that fluting had everywhere disappeared by the beginning of the Holo- cene. Probably, its apogee occurred during the El Abra Stadial (Politis, 1991, p. 297). Fluted points and life styles No published fluted points have been associated with faunal or macrobotanical remains. Jaymes (pers. comm., 1992), however, has found 'Fish-Tail' stemmed points associated with extinct megafauna in Falcon state (Venezuela) (Fig. 5h). Elsewhere, extinct mammals are nearly always present where fluted points accom- pany animal remains, e.g. mammoth (Mammuthus primigenius) in the southwestern United States, and horse, ground sloth, and glyptodontids in Argentina and Chile. If these taxa vanished before the end of the LGS (despite Politis et al., 1987), their extinction and the disappearance of vertical basal fluting may not be coincidental. Nevertheless, the consensus opinion among archaeologists is that consuming extinct her- bivores does not necessarily equate with hunting 'specialization'. Opportunistic selection for the most abundant and/or easiest-to-hunt taxa in particular habitats, be they now extinct or extant, is a more likely scenario. Guanaco were important prey species for fluted-point users in Patagonia and the pampas; white- tailed deer probably were in this study area. At the same time, even if 'Clovis' weapons were not effective against adult mammoth (Prison, 1989) and even if underutilized megamammal carcasses were scavenged (Haynes, 1991), fluted points and the geographically uniform scrapers, perforators, and burins that accom- pany them all over the American continent (see Figs 3i, k, 1 and 4g, h, k-n) were not designed for scavenging! If we adjudicate fluted point finds to the 11th millen- nium BP, we find their users in vegetation types that can be inferred from palaeoecological data as (1) paramo (highland Guatemala), (2) montane oak forests (drier in central Panama than Caribbean Costa Rica?), (3) seasonal lowland forest (Caribbean central Panama (Lake Gatun), Gulf of Uraba), (4) dry tropical forest, perhaps thorn-scrub and wooded savanna (central Panamanian lowlands, parts of Venezuela, Magdalena lowlands), (5) Andean oak forests (Cauca valley), and (6) semi-desert (Paraguana Peninsula, Venezuela). A sudden increase in particulate charcoal at Lake La Yeguada (Panama) after ca. 11,000 BP suggests that humans began to burn fires in oak-//ex-magnolia forests at this time. Forest plant products and small animals were surely collected (Clausen et al., 1979; Borrero, 1995; Lynch, 1983; Piperno et al., 1991a). Nevertheless, while the plant resources of the above 186 R. Cooke habitats would have been diverse (requiring different foraging strategies and preparation technology), the trophic behavior of deer, sloth, horse, and mastodonts did not restrict these taxa to specific vegetation types. All fluted point finds in the study area come from areas where seasonal water shortage (and animal aggrega- tion behavior) was surely accentuated in YD-times. Lakes would have been primary foci for prey species. If the search for dietarily (and probably cognitively) important prey species was one motivation for fluted- point users to habitually cross ecological boundaries, another was the desire for the best raw material for bifacial flaking. At tool manufacturing or repair sites (Los Tapiales, La Mula-West, La Elvira) lithic raw materials were brought in from afar. The local rarity of good quality stone may explain why bifacial points, carefully trimmed scrapers, burins, and perforators are rare or absent in Sabana de Bogota LGS (Correal Urrego and van der Hammen, 1977). Hence, the 'Abriense'-'Tequendamiense' dichotomy could reflect LGS hunter-gatherer behavior, rather than mutually exclusive tool traditions. Conversely, the large size of the Finca Guardiria workshop in Costa Rica (100,000 m2) is perhaps related to the abundance of good quality stone within the site's limits (Snarskis, 1979). Holocene environmental change and human subsistence A YD-related cooling episode was identified in the Costa Rican La Chonta bog between ca. 11,100 and 10,400 BP. In the Colombian Andes, the El Abra Sta- dial lasted from ca. 10,900-10,150 BP. At La Yeguada (Panama) the period 11,000-10,300 BP represents 'a major boundary in all the (palaeoecological) records' (Bush et al, 1992, p. 259). The La Yeguada lake core is the only one from the tropical lowlands where pre-8000 BP human activities are clearly visible in the pollen and phytolith record. The burning episode just mentioned was followed (10,500-8000 BP) by an increased abundance of grass and disturbance-indicator plants (Piperno et al., 1990, 1991b). Archaeological evidence from three rock- shelters located between this lake and the Pacific coast (Corona, Carabali, and Vampiros) substantiates (1) continued human presence after climatic amelior- ation, (2) wild plant food collection in seasonally dry forests, and (3) the use of a cultigen (arrowroot, Maranta arundinacea). Bifacial flaking debris in depo- sits dated between 10,400 + 450 and 8040 ? 390 BP accompanies arrowroot microremains and edge- ground cobbles (for mashing the tubers?) (Cooke, 1993; Cooke and Ranere, 1992a). No bifacial point fragments were found in these excavations; but it is likely that lipped thinning flakes were removed from square- stemmed barbed points similar to surface-collected examples from the La Mula-Central quarry (Ranere and Cooke, 1995: Fig. 1.3). Surface-collected non-fluted bifacial projectile points and turtle-backed scrapers (cf. Fig. 3(1)) from Costa Rica, Guatemala, and Belize probably include pre-8000 BP materials (Acuha Coto, 1983; Brown, 1980; Kelly, 1993). Sound contextual information for the contem- poraneity of unfluted bifacial tools made of cryptocrys- talline stone, forest plant products, and their processing tools comes from San Isidro, in the valley of Popayan (altitude: 1800 m). Gnecco Valencia (1994) identifies this single component site, anchored to two dates (10,050 + 100 and 9530 + 100 BP), as a knapping sta- tion where hunter-foragers made artifacts and gathered mostly palm fruits (cf. Acrocomia), but also Virola, Caryocar, and Persea (bone was not recovered). Cobble grinding tools, smoothed discoidal stones (Fig. 3g), and a smoothed axe (Fig. 3h) were stratified with them. Projectile points include foliate, double-triangular, stemmed, and lanceolate types, all unfluted (Fig. 5c and d). Large bifaces (Fig. 3j), carefully prepared end- scrapers, and edge-trimmed unifacial tools are also present. Gnecco Valencia argues that San Isidro and La Elvira, which show clear cultural affiliations with adjacent parts of Ecuador (Bell, 1965; Mayer-Oakes, 1986), overlapped chronologically; hence, their diver- gent lithic traditions allude to coeval cultural hetero- geneity. It remains to be seen whether this hypothesis is stronger than its obvious alternate (short-lived dia- chronic changes in widely distributed tool types). Hunter-gatherer cultural diversity after ca. 10,500 BP has also been proposed by Lopez Castano for the middle Magdalena lowlands. Stone tools and flaking debris are widely distributed in an area where excellent cryptocrystalline stone abounds. Particularly wide- spread are unifacial ovate turtle-backed scrapers (e.g. Fig. 3(1)) and large flakes with carefully trimmed edges (e.g. Fig. 3k). The latter was found at the single- component La Palestina site associated with thinning flakes, flake scrapers, cores, and flaking debris (with two dates: 10,400 ? 90 and 10,230 + 80 BP). Neither this site, nor coeval San Juan de Bedout (10,350 + 90 BP), provided stratified examples of the wide variety of stemmed and tanged points that were collected from the surface (Lopez Castano, 1995a, b; Oleoducto de Colombia, 1994). Lopez Castano (1993, 1995a, b) infers from these data that (1) several late LGS/Early Holocene lithic traditions existed (some of these coeval), (2) hunter- gatherer groups with a post-Palaeoindian biface- scraper industry survived locally until 5000-3000 BP, and (3) an itinerant, savanna-adapted way-of-life per- sisted through the Early Holocene. Hunter-gatherer societies that used bifacial projectile points and empha- sized deer and/or camelid hunting were present in the Ecuadoran and Peruvian Andes until well after 8000 BP (Rick, 1980; Lynch, 1980; Lynch and Pollock, 1981). More excavations at single-component sites may, or may not, confirm the hypothesis of diachronically intensifying hunter-gatherer cultural diversity in Colombia (some undated points (e.g. Fig. 5i from Human Settlement of Central America and Northernmost South America 187 Yondo, Magdalena river) are clearly pressure-flaked, unlike other LGS examples that this author has han- dled). The utilization of Andean forest products, appar- ent at San Isidro, is substantiated for the Early Holocene in the Calima Valley. At Sauzalito (with two dates: 9670 + 170 and 9600+110 BP), a flaked laterally notched tool of igneous material (Fig. 3c) has been interpreted as a 'hoe'. At nearby El Recreo, a well- smoothed 'hoe' (Fig. 3d) was associated with three dates with a Iff range of 8910-7690 BP. The wide distribution of these 'hoes' throughout the middle and upper Cauca valley (Salgado Lopez, 1989), as well as their stratigraphic association with cobblestone grinders, and carbonized palm nuts and Persea pits, points towards a regional sylvan lifestyle at the Andean/Sub-Andean forest boundary. A single thinn- ing flake at Sauzalito is the only evidence at these sites for the continuation of LGS bifacial reduction pro- cesses. The bifacial reduction of tools made of crypto- crystalline stone seems gradually to have petered out in the Early Holocene, as it did in central Panama (Ranere, 1992). The record for human activities in the Sabana de Bogota suggests continuing, generally more intensive, but still intermittent occupation of rockshelters used in the LGS, and the incorporation of higher elevations into the subsistence round. Structures have been identi- fied at one open-site (Checua). Neusa II (altitude: 3350 m) was probably located at the upper limit of Andean forest when it was first occupied. The lithic industry from DU2 (8370 + 90 BP) is 'Abriense' (tools are identified as 'scrapers' and 'knives'). DU 2A, im- mediately above D2 inside the shelter, contains anvils, grinding bases, and edge-ground cobbles similar to those reported from the Carabali shelter in Panama (8040 +90 BP: Ranere and Cooke, 1995 (p. 13)). White-tailed and brocket deer (Mazama), armadillo, rodents, and bird were taken. (If correctly identified, the black-collared hawk (Busarellus nigricollis) was none- theless probably not locally available) (Rivera Escobar, 1992). Compositionally and artifactually the Early Holo- cene deposits at Tequendama and El Abra are similar. DU 7 at Tequendama II (estimated non-radiometric age: 9500-8500 BP) contains animals remains clus- tered around clay-lined hearths. By this time, the site was within Andean forest. 'Keeled' scrapers are present, but bifacial artifacts are not. Possible lithic innovations are hook-shaped flake 'perforators' and 'spokeshaves', the latter perhaps related to the increased importance of bone tools. Guinea-pig displaced the white-tailed deer as favorite preferred prey species, both here and at El Abra, where DU D (ca. 9400-7250 BP) is judged to be coeval with Tequendama DU 7 (Correal Urrego and van der Hammen, 1977; van der Hammen, 1978). Nemocon IV's varied fauna and sparse occupation debris allude to its use as a hunting camp ca. 9000-8000 BP. The predominance of white-tailed deer, and the presence of guinea-pigs, fox (Urocyon), pecca- ries (Tayassuidae), and howler monkey (Alouatta) sug- gest either an ecotonal (forest/open country) location or a wide hunting range. Land-snails had been added to the sabana diet by this time (Correal Urrego, 1979). The existence of open-air living sites in the sabana at the beginning of the Holocene is demonstrated by the Galindo and Checua sites. The earliest low-density occupation at Checua included post-holes and hearths associated with a lone date of 8200 + 110 BP. Edge- ground cobbles and other pounding instruments allude to the plant food processing. This site and Nemocon 4 may have been used seasonally by the same cultural group (Groot de Mahecha, 1995). CONCLUDING REMARKS This review underlines some high-probability scen- arios that can be inferred from the summarized data, some alternate scenarios whose resolution awaits better-quality data, and some areas of investigation whose encouragement would improve knowledge of the elected topic. High probability scenarios: (1) A small and widely dispersed pre-El Abra Stadial (YD) human population. (2) The pre-El Abra Stadial (YD) consumption of ex- tinct herbivores, including horse and gompho- theres, by cultural groups who made bifacial lanceolate points (probably socketed into wooden spears) and carefully trimmed scrapers. (3) More or less coincident with the YD and El Abra Stadial (10,500 + 600 BP), a widely distributed, cul- turally affiliated, and highly mobile ('Palaeoin- dian') population that customarily utilized basally fluted bifacial artifacts, blades, burins, flake perfor- ators, and carefully trimmed end- and side-scrapers (often 'turtle-backed'). (4) The use of different forest and open vegetation types throughout the LGS from sea level up to ca. 2500 m (except non-seasonal lowland rainforest). (5) At or soon after the LGS/Holocene boundary, widespread forest-dwelling settlements, whose in- habitants habitually collected forest plant products, some of which appear to have been under domesti- cation by 8000 BP. Alternative scenarios which cannot be confirmed by current data: (1) Humans were/were not present during the Upper Pleniglacial and did/did not use extinct Pleistocene mammals such as mastodons. (2) All pre-El Abra Stadial populations were/were not acquainted with bifacial reduction processes. (3) Pre-Holocene populations did/did not habitually collect wild plant foods. (4) The extinction of Pleistocene mammals did/did not coincide chronologically with the disappearance of fluted points. (5) Extant fauna, such as white-tailed deer, were/were not an important LGS hunting resource. 188 R. Cooke (6) LGS and Early Holocene tool kits represent either short-lived, but areally distributed artifact types, or coeval assemblages belonging to different cultural groups with distinct subsistence orientations. (7) Hunter-gatherers using bifacial projectile points and large blade/flake scrapers did/did not survive after 8000 BP coevally with incipient cultivators. (8) Universal forest expansion and/or increased seden- tism after climatic amelioration did/did not lead to the development of several economic systems strongly correlated with specific geographic regions and/or vegetation types (i.e. an 'Andean hunting tradition', 'a Central American premontane horti- cultural system)'. Areas of study that should be promoted: (1) Trace-wear analysis (cf. Nieuwenhuis, 1994; Gnecco Valencia and Mohammed, 1994). (2) Studies of lithic waste materials (for identifying bifacial thinning and fluting flakes, and lithic reduc- tion processes, especially where bifacial 'tools' have not been reported. (3) Excavation at buried single-component sites (to compensate for dating inconsistencies at contin- ually occupied sites). (4) Archaeology and palaeoecology in evergreen (humid) rainforests. ACKNOWLEDGEMENTS I am deeply indebted to the following institutions and colleagues for expeditiously sending me bibliographic materials, for allowing me to make use of unpublished data, for letting me take photographs of stone tools, and for sending me slides: Instituto Colombiano de Antropologia, Gonzalo Correal Urrego, Thomas van der Hammen, Channah Nieuwenhuis, Carlos Lopez Castano, Cristobal Gnecco Valencia, Dolores Piperno, and Jose Oliver. Alejandro Caballero composed the illustrations at the Smithsonian Tropical Research Institute's Digital Processing Laboratory in Panama City. I also express my gratitude to Cristobal Gnecco Valencia, German Peha Leon, Marianne Cardale de Schrimpff, and Ana Maria Falchetti for special attentions during recent visits to Colombia to gather informa- tion for this synthesis, whose omissions, and errors of interpretation and fact are mine alone. REFERENCES Acuna Coto, V. (1983). Florencia-1, un sitio preceramico en la ver- tiente atlantica de Costa Rica. V'inculos, 9, 1-14. Ardila Calderon, G (1991). The peopling of Northern South Amer- ica. In: Bonnichsen, R. and Turnmire, K.L. (eds), Clovis: Origins and Adaptations, pp. 261-282. Center for the Study of the First Americans, Orvallis, OR. Ardila Calderon, G.I. and Politis, G. (1989). Nuevos datos para un viejo problema: investigation y discusiones en torno del pob- lamiento de America del Sur. Museo del Oro, Boletin, 23, 3-45. Bell, RE. (1965). Investigaciones Arqueologicas en el Sitio de El Inga, Ecuador. Editorial Casa de la Cultura Ecuatoriana, Quito. Bird, J.B. (1969). A comparison of South Chilean and Ecuadorean 'Fish-Tail' projectile points. The Kroeber Anthropological Society Papers, 40, 52-71. Bird, J.B. (1988). Travels and Archaeology in South Chile (edited by J. Hyslop, with journal segments by Margaret Bird). University of Iowa Press, Iowa City. Bird, J.B. and Cooke, R.G (1977). Los artefactos mas antiguos de Panama. Revista Nacional de Cultura, 6, 7-31. Bird, J.B. and Cooke, R.G. (1978). The occurrence in Panama of two types of Paleo-Indian projectile point. In: Bryan, A.L. (ed.), Early Man in America from a Circum-Pacific Perspective, pp. 263-272. Occasional Papers No. 1, Department of Anthropology, Univer- sity of Alberta, Edmonton. Borrero, L.A. (1995). The archaeology of the far south of America ? Patagonia and Tierra del Fuego. In: Johnson, E. (ed.), Ancient Peoples and Landscapes, pp. 207-215. Museum of Texas Tech University, Lubbock. Botiva Contreras, A. (n.d.). La Altiplanice Cundiboyacense. In: Colombia Prehispanica: Regiones Arqueologicas, pp. 77-115. Colcultura/ICAN, Bogota. Bradbury, J.P., Leyden, B., Salgado-Labouriau, M., Lewis Jr, W.M., Schubert, C, Binford, M.W., Frey, D.G, Whitehead, DR. and Weizbahn, F.H. (1981). Late Quaternary environmental history of Lake Valencia, Venezuela. Science, 214, 1299-1305. Brenner, M. (1993). Lakes Salpeten and Quexil, Guatemala, Central America. In: Gierolowski, E. and Kelts, H. (eds), Global Geological Record of Lake Basins, Vol. 1, pp. 377-380. Cambridge University Press, Cambridge. Brown, K.L. (1980). A brief report on Palaeoindian-Archaic occupa- tion in the Quiche Basin, Guatemala. American Antiquity, 45, 313-324. Bryan, A.L. (1991). The fluted point tradition in the Americas ? one of several adaptations to the Late Pleistocene American environ- ments. In: Bonnichsen, R. and Turnmire, K.L. (eds), Clovis: Origins and Adaptations, pp. 15-34. Center for the Study of the First Americans, Orvallis, OR. Bryan, A.L., Casamiquela, R.M., Cruxent, J.M., Gruhn, R. and Oschenius, C. (1978). An El Jobo mastodon kill at Taima- taima, Venezuela. Science, 200, 1275-1277. Bryan, A.L. and Gruhn, R. (1979). The radiocarbon dates of Taima- Taima. In: Oschenius, C. and Gruhn, R. (eds), Taima-Taima, a Late Pleistocene Kill-site in Northernmost South America, pp. 53-58. South American Quaternary Documentation Program. Bush, M.B. and Colinvaux, P. A. (1990). A pollen record of a complete glacial cycle from lowland Panama. Journal of Vegetation Science, 1, 105-8. Bush, MB., Piperno, DR., Colinvaux, PA., de Oliveira, P.E., Krissek, L.A, Miller, M.C. and Rowe, W.E. (1992). A 14,300-yr paleoecological profile of a lowland tropical lake in Panama. Ecological Monographs, 62, 251-275. Clausen, C.J., Cohen, AD, Emiliani, C, Holman, J.A. and Stipp, J.J. (1979). Little Salt Spring, Florida: a unique underwater site. Science, 203, 609-614. Cooke, R.G. (1993). Eta pas tempranas de la produccion de alimentos vegetales en la Baja Centroamerica y partes de Colombia (Region Historica Chibcha-Choco). Revista de Arqueolog'ia Americana, 6, 35-70. Cooke, R.G. and Ranere, A.J. (1992a). Prehistoric human adapta- tions to the seasonally dry forests of Panama. World Archaeology. 24, 114-133. Cooke, R.G. and Ranere, A.J. (1992b). The origin of wealth and hierarchy in the Central Region of Panama (12,000-2000 BP), with observations on its relevance to the history and phytogeny of Chibchan-speaking polities in Panama and elsewhere. In: Lange, F. (ed.), Wealth and Hierarchy in the Intermediate Area, pp. 243-316. Dumbarton Oaks, Washington, DC. Correal Urrego, G (1979). Investigaciones Arqueologicas en Abrigos Rocosos de Nemocon y Sueva. Fundacion de Investigaciones Arqueologicas Nacionales (Banco de la Republica), Bogota. Correal Urrego, G. (1981). Evidencias Culturales y Megafauna Pleis- tocenica en Colombia. Fundacion de Investigaciones Arque- ologicas Nacionales, Banco de la Republica. Litografia Arco, Santafe de Bogota. Correal Urrego, G. (1986). Apuntes sobre el medio ambiente Pleis- tocenico y el hombre prehistorico en Colombia. In: Bryan, A.L. (ed.), New Evidence for the Pleistocene Peopling of the Americas, pp. 115-131. Center for the Study of Early Man, University of Maine, Orono. Correal Urrego, G. (1993). Nuevas evidencias culturales pleis- tocenicas y megafauna en Colombia. Boletin de Arqueolog'ia, 8, 3-13. Correal Urrego, G and van der Hammen, T. (1977). Investigaciones Arqueologicas en los Abrigos Rocosos del Tequendama: 12,000 Anos de Historia del Hombre y de su Medio Ambiente en la Altiplanicie de Bogota. Biblioteca Banco Popular, Bogota. Correal Urrego, G, van der Hammen, T. and Lerman, J.C. (1966-69). Artefactos liticos de abrigos rocosos en El Abra, Colombia: informe prelininar. Revista Colombiana de Arqueolog'ia, 14, 12-46. Human Settlement of Central America and Northernmost South America 189 Correal Urrego, G. and Pinto Nolla, M. (1993). Evidencia de caza- dores especializados en el sitio de La Gloria, Golfo de Uraba. Appendix to: Investigation Arqueologia en el Municipio de Zipa- cbn, Cundimarca, pp. 195-205. Fundacion de Investigaciones Arqueologicas Nacionales, Banco de la Republica, Bogota. Cruxent, J.M. (1979). Stone and bone artifacts from Taima-Taima. In: Oschenius, C. and Gruhn, R. (eds), Taima-Taima, a Late Pleis- tocene Kill-site in northernmost South America, pp. 77-92. South American Quaternary Documentation Program. Denton, G.H. and Hendy, C.H. (1994). Younger Dryas advance of Franz Josef glacier in the southern Alps of New Zealand. Science, 264, 1434-1437. Dillehay, T.D., Ardila Calderon, G, Politis, G. and Coutinho Belt- rao, M. da C. (1992). Earliest hunters and gatherers of South America. Journal of World Prehistory, 6, 145-204. Dillehay, T.D. and Collins, M.B. (1991). Monte Verde, Chile: a com- ment on Lynch. American Antiquity, 56, 333-341. Dunbar, J.S. (1991). Resource orientation of Clovis and Suwanee Paleoindian sites in Florida. In: Bonnichsen, R. and Turnmire, K.L. (eds), Clovis: Origins and Adaptations, pp. 185-213. Center for the Study of the First Americans, Orvallis, OR. Frison, G.C (1989). Experimental use of Clovis weaponry and tools on African elephants. American Antiquity, 54, 766-784. Gnecco Valencia, C. (1994). The Pleistocene/Holocene boundary in the Northern Andes: an Archaeological Perspective. Ph.D. Disser- tation, Washington University, St. Louis, unpublished. Gnecco Valencia, C. and Mohammed, A. (1994). Tecnologia de cazadores-recolectores subandinos: analisis funcional y organization tecnologica. Revista Colombiana de Antropologia, 31, 7-31. Gnecco Valencia, C. and Salgado Lopez, H. (1989). Adaptaciones Preceramicas en el Suroccidente de Colombia. Museo del Oro, Boletin, 24, 35-53. Groot de Mahecha, A.M. (1995). Checua: un aporte para el cono- cimiento del preceramico de la Sabana de Bogota. In: Cavelier, I. and Mora, S. (eds), Amhito y Ocupaciones Tempranas de la America Tropical, pp. 45-58. Fundacion Erigaiae/ICAN, Santafe de Bogota. Gruhn, R. and Bryan, A.L. (1977). Los Tapiales: a Paleo-Indian campsite in the Guatemalan highlands. Proceedings of the Ameri- can Philosophical Society, 121, 235-273. Gruhn, R. and Bryan, A.L. (1984). The record of Pleistocene mega- faunal extinction at Taima-Taima, northern Venezuela. In: Martin, PS. and Klein, R.G. (eds), Quaternary Extinctions: a Pre- historic Revolution, pp. 128-137. University of Arizona Press, Tuscon. Gruhn, R. and Bryan, A.L. (1991). A review of Lynch's descriptions of South American Pleistocene sites. American Antiquity, 56, 342-348. Hayden, B. (1980). A fluted point from the Guatemalan highlands. Current Anthropology, 21, 702. Haynes, C.V. (1991). Archaeological and paleohydrological evidence for a terminal Pleistocene drought in North American and its bearing on Pleistocene extinction. Quaternary Research, 35, 438-450. Haynes, G (1991). Mammoths, Mastodonts, and Elephants: Biology, Behavior, and the Fossil Record. Cambridge University Press, Cambridge. Herrera, L., Bray, W., Cardale de Schrimpff, M. and Botero, P. (1992). Nuevas fechas de radiocarbono para el Preceramico en la Cordill- era Occidental de Colombia. In: Ortiz-Tronoco, O. and Hammen, T. van der (eds), Archaeology and Environment in Latin America, pp. 145-164. Universiteit van Amsterdam, Amsterdam. Hooghiemstra, H. and van der Hammen, T. (1993). Late Quaternary vegetation history and paleoecology of Laguna Pedro Palo (sub- andean forest belt, Eastern Cordillera, Colombia). Review of Palaeohotany and Palynology, 77, 235-262. Horn, S. (1990). Timing of deglaciation in the Cordillera de Talamanca. Climate Research, 1, 81-83. Hurt, W., van der Hammen, T and Gonzalo Urrego, G (1972). Preceramic sequences in the El Abra rock-shelters, Colombia. Science, 175, 1106-1108. Hurt, W.R., van der Hammen, T. and Gonzalo Urrego, G (1977). The El Abra rockshelters, Sabana de Bogota, Colombia, South America. Indiana University Occasional Papers and Monographs, 2. Islebe, G.A., Hooghiemstra, H. and van der Borg, K. (1995). A cool- ing event in the Younger Dryas Chron in Costa Rica. Palaeogeo- graphy, Palaeoclimatolog y, Palaeoecology, 117, 73-80. Kelly, T.C. (1993). Preceramic projectile-point typology in Belize. Ancient Mesoamerica, 4, 205-227. Leyden, B.W. (1985). Late Quaternary aridity and Holocene moist- ure fluctuations in the Lake Valencia basin, Venezuela. Ecology, 66, 1279-1285. Leyden, B.W., Brenner, M., Hodell, DA. and Curtis, J.A. (1993). Late Pleistocene climate in the Central American lowlands. Climate Change in Continental Isotopic Records, Geophysical Monograph, 28, 165-178. Lopez Castano, CE. (1993). La edad y el ambiente preceramico en el Magdalena Medio: resultados de laboratorio del sitio Penones de Bogota. Boletin de Arqueologia, Fundacion de Investigaciones Arqueologicas Nacionales, 8, 13-25. Lopez Castano, CE. (1995a). Dispersion de puntas de proyectil bifaciales en la cuenca media del rio Magdalena. In: Cavelier, I. and Mora, S. (eds), Amhito y Ocupaciones Tempranas de la America Tropical, pp. 73-82. Fundacion Erigaie and Instituto Colom- biano de Antropologia, Santafe de Bogota. Lopez Castano, CE. (1995b). Preceramic hunter-gatherers in the tropical lowlands of the Middle Magdalena valley (Colombia, South America). Paper Presented at the 60th Annual Meeting of the Society for American Archaeology, Minneapolis. Lowell, T.V., Heusser, C.J., Andersen, B.G., Moreno, PL, Hauser, A., Heusser, L.E., Schliichter, C, Marchant, DR. and Denton, G.H. (1995). Interhemispheric correlation of Late Pleistocene glacial events. Science, 269, 1541-1549. Lynch, T.F. (ed.) (1980). Guitarrero Cave: Early Man in the Andes. Academic Press, New York. Lynch, T.F. (1983). The Paleo-Indians. In: Jennings, J.D. (ed.), Ancient South Americans, pp. 87-137. Freeman, San Francisco. Lynch, T.F. (1990). Glacial-age man in South America? A critical review. American Antiquity, 55, 12-36. Lynch, T.F. and Pollock, S. (1981). La arqueologia de la Cueva Negra de Chobshi. Misceldnea Antropolbgica Ecuatoriana, 1, 92-119. Markgraf, V. (1989). Palaeoclimates in Central and South America since 18,000 BP based on pollen and lake-level records. Quater- nary Science Reviews, 8, 1 -24. Mayer-Oakes, W.J. (1986). El Inga: a Paleo-Indian site in the sierra of northern Ecuador. Transactions of the American Philosophical Society, 76(4), 1-235. Meltzer, M.J. (1995). Clocking the first Americans. Annual Reviews in Anthropology, 24, 21-45. Morlan, RE. (1988). Pre-Clovis people: early discoveries of America? In: Carlise, R.C (ed.), Americans before Columbus: Ice-Age Origins, pp. 31-43. Ethnology Monographs, 12, Department of Anthro- pology, University of Pittsburgh. Nieuwenhuis, Ch. (1994). Use-wear analysis on Preceramic Colom- bian artefacts of the Abriense toolclass. Praehistoria Leidensia, 26, 1-7. Oleoducto de Colombia (1994). Arqueologia de Rescate. Oleoducto Vasconia-Covehas/Un Viaje por el Tiempo a lo Largo del Ole- oducto/Cazadores-Recolectores, Agroalfareros y Orfebres. Santafe de Bogota, 1994. Oliver, JR. and Alexander, CS. (1990). The Pleistocene peoples of western Venezuela: the terrace sequence of rio Pedregal and new discoveries in Paraguana. Paper Presented at the First World Summit on the Peopling of the Americas (Center for the Study of the First Americans-University of Maine), Orono, Maine, 1989. Oschenius, C. (1979). Paleoecology of Taima-Taima and its sur- roundings. In: Oschenius, C. and Gruhn, R. (eds), Taima-Taima, a Late Pleistocene Kill-site in northernmost South America, pp. 91-104. South American Quaternary Documentation Program. Piperno, DR. (1995). Late Pleistocene early Holocene human ecol- ogy of central America. Paper Presented at the 60th Annual Meeting of the Society for American Archaeology, Minneapolis. Piperno, DR., Bush, M.B. and Colinvaux, P.A. (1990). Paleoenviron- ments and human occupation in Late-Glacial Panama. Quater- nary Research, 33, 108-116. Piperno, DR., Bush, M.B. and Colinvaux, P.A. (1991a). Paleoeco- logical perspectives on human adaptation in Panama. I. The Pleistocene. Geoarchaeology, 6, 201-225. Piperno, DR., Bush, M.B. and Colinvaux, P.A. (1991b). Paleoeco- logical perspectives on human adaptation in Panama. II. The Holocene. Geoarchaeology, 6, 227-250. Politis, GG (1991). Fishtail points in the Southern Cone of South America: an Overview. In: Bonnichsen, R. and Turnmire, K.L. (eds), Clovis: Origins and Adaptations, pp. 287-301. Center for the Study of the First Americans, Orvallis, OR. Politis, G.G., Tonni, E., Fidalgo, F., Salemme, M. and Guzman, L.M. (1987). Man and Pleistocene megamammals in the Argentine 190 R. Cooke Pampa: site 2 at Arroyo Seco. In: Current Research in the Pleistocene, pp. 159-162. Center for the Study of Early Man, Orono. Ranere, A.J. (1992). Implements of change in the Holocene environ- ments of Panama. In: Ortiz Troncoso, OR. and Hammen, T. van der (eds), Archaeology and Environment in Latin America, pp. 25-44. Universiteit van Amsterdam, Amsterdam. Ranere, A.J. and Cooke, R.G. (1991). Paleo-Indian occupation in the Central American Tropics. In: Bonnichsen, R. and Turnmire, K.L. (eds), Clovis: Origins and Adaptations, pp. 237-253. Center for the Study of the First Americans, Corvallis, Oregon. Ranere, A.J. and Cooke, R.G. (1995). Evidencis de ocupacion humana en Panama a postrimerias del Pleistoceno y comienzos del Holocene. In: Cavelier, I. and Mora, S. (eds), Ambito y Ocu- paciones de la America Tropical, pp. 5-26. ICAN/Colcultura/ Fundacion Erigaie, Santafe de Bogota. Reichel-Dalmatoff, G. (1965). Colombia. Thames and Hudson, London. Reichel-Dalmatoff, G. (1992). Colombia Indigena: Periodo Precolombino. In: Manual de Historia de Colombia, Tomo 1, 4th edn, pp. 33-115. Procultura S.A., Santa fe de Bogota. Rick, J. (1980). Prehistoric Hunters of the High Andes. Academic Press, New York. Rivera Escobar, S. (1992). Neusa: 9,000 afios de Presencia Humana en el Paramo. Fundacion de Investigaciones Arqueologicas Nacionales, Banco de la Republica, Santafe de Bogota. Rouse, I. and Cruxent, J.M*. (1963). Venezuelan Archaeology. Yale University Press, New Haven, CT. Salgado Lopez, H. (1989). Medio Ambiente y Asentamientos Humanos Prehispanicos en el Calima Medio. Instituto Vallecaucano de Investigaciones Cientificas, Call. Schreve-Brinkman, E.J. (1978). A palynological study of the Upper Quaternary sequence in the El Abra corridor and rockshelters (Colombia). Palaeogeography, Palaeoclimatolog y, Palaeoecology, 25, 1-109. Schubert, C. (1988). Climatic changes during the last glacial max- imum in Northern South America and the Caribbean: a review. Interciencia, 13, 128-137. Sheets, P.O. (1994). Clipped stone artifacts from the Cordillera de Tilaran. In: Sheets, P.D. and McKee, B.R. (eds), Archaeology, Sedentism, and Remote Sensing in the Arenal Region, Costa Rica, pp. 211-254. University of Texas Press, Austin. Snarskis, M.J. (1979). Turrialba: a Paleo-Indian quarry and work- shop site in eastern Costa Rica. American Antiquity, 44, 126-137. Uribe, M.V. (1990). Cronologia absoluta de la arqueologia colom- biana. Revista de Antropologia y Arqueologia, 6, 207-233. van der Hammen, T. (1978). Stratigraphy and environments of the Upper Quaternary of the El Abra corridor and rockshelters (Colombia). Palaeogeography, Palaeoclimatology, Palaeoecology, 25, 111-162. van der Hammen, T. (1992). Historia, Ecologia y Vegetacibn. FEN/ COA/FPCBP, Santafe de Bogota. van der Hammen, T. and Absy, M.L. (1994). Amazonia during the last glacial. Palaeogeography, Palaeoclimatology, Palaeoecology, 109, 247-261. van der Hammen, T. and Correal-Urrego, G. (1992). El hombre prehistorico en la Sabana de Bogota: datos para una prehistoria ecologica. In: Hammen, T. van der (ed.), Historia, Ecologia y Vegetacibn, pp. 217-231. FEN/COA/FPCBP, Santafe de Bogota. van der Hammen, T. and Gonzalez, E. (1960). Upper Pleistocene and Holocene climate and vegetation of the Sabana de Bogota (Colombia, South America). Geologie en Mijnbouw, 39, 737-746. van der Hammen, T. and Hooghiemstra, H. (1995). The El Abra stadial, a Younger Dry as equivalent in Colombia. Quaternary Science Reviews, 14, 841-852.