Search

It has been proposed that a multiregional model could describe how Homo sapiens evolved in Africa beginning 300,000 years ago. Multiregionalism would require enduring morphological or behavioral differences among African regions and morphological or behavioral continuity within each. African fossils, archeology, and genetics do not comply with either requirement and are unlikely to, because climatic change periodically disrupted continuity and reshuffled populations. As an alternative to multiregionalism, I suggest that reshuffling produced novel gene constellations, including one in which the additive or cumulative effect of newly associated genes enhanced cognitive or communicative potential. Eventual fixation of such a constellation in the lineage leading to modern H. sapiens would explain the abrupt appearance of the African Later Stone Age 50-45 thousand years ago, its nearly simultaneous expansion to Eurasia in the form of the Upper Paleolithic, and the ability of fully modern Upper Paleolithic people to swamp or replace non-modern Eurasians.

Klein (1979, 1994) reported a contrast in ungulate species frequencies between the Middle Stone Age (MSA) layers of Klasies River Main (KRM) and the Later Stone Age (LSA) layers of nearby Nelson Bay Cave (NBC), South Africa. Early European travelers noted that the distinctive subtropical broadleaf forest surrounding both coastal sites housed exceptional numbers of buffalo (Syncerus caffer) and bushpig (Potamochoerus larvatus; Skead, 1980). They failed to mention eland, suggesting that eland were locally absent or rare, in keeping with this species’ tendency to avoid forest (Skinner and Chimimba, 2005). At both KRM and NBC the deposits formed under similar climatic conditions (Last Interglacial and the Holocene Interglacial, respectively), yet only the NBC LSA fauna anticipated the historic forest fauna in the abundance of buffalo and bushpig and the rarity of eland. The unanticipated abundance of eland in the KRM MSA fauna led Klein to suggest that the KRM people were less competent hunters. Unlike buffalo and bushpig, which commonly counterattack predators, eland tend to flee, and associated artifacts suggest that only the NBC LSA hunters had projectile weapons that would have allowed them to hunt dangerous species from a relatively safe distance. The NBC LSA people could then have obtained buffalo and bushpig more in proportion to their live abundance. Lacking projectiles, the KRM MSA people would have tended to capture more of the less threatening eland, which they could have driven over nearby cliffs. Driving could explain why the KRM eland were mainly prime-age adults, whereas the KRM buffalo were mostly very young or old individuals, as would be expected if the hunters focused on the most vulnerable buffalo or acquired them by scavenging. Conceivably, culture was irrelevant, and topography or some other strictly local factor explains the abundance of eland at KRM. However, a test for this would require additional fossiliferous MSA sites in the same forest, and KRM remains the only one. The KRM/ NBC contrast is roughly replicated between Die Kelders Cave 1 (DK1) and nearby Byneskranskop 1 (BNK1), where eland dominate buffalo and bushpig in the DK1MSA layers and buffalo and bushpig dominate eland in the BNK1 LSA layers (Klein and Cruz-Uribe, 1996). However, DK1 and BNK1 are located 400e500 km east of KRM and NBC (Fig. 1) in a scrub/thicket mosaic where buffalo and bushpig may not have outnumbered eland historically. In addition, the DK1 and BNK1 faunas accumulated under differing climatic conditions (mainly glacial and interglacial, respectively), and a paleoenvironmental difference could thus explain the faunal contrast. The DK1/BNK1 contrast underscores the need to control for environment before invoking cultural or human behavioral explanations for differences between sites. With Klein’s interpretation of the KRM/NBC contrast in mind, Faith (2008) undertook log-transformed regressions of eland and buffaloþ pig abundance on the total number of ungulates in 95 LSA samples and 27 MSA samples from Lesotho and South Africa. He concluded that “eland, buffalo, andwild pig are equally abundant in the MSA and LSA” (p. 24) and therefore that MSA and LSA people were equally adept hunters. We show below that Faith’s results are based on inappropriate use of regression, but even if we were to accept them, they do not show that MSA and LSA people were behaving in the same way. Eland, buffalo, and wild pigs could be equally abundant, on average, in MSA and LSA sites only if their live proportions (relative numbers) were usually the same nearby, or if MSA and LSA people, faced with different proportions at different times and places, managed to equalize the proportions in their sites. Faith’s result could mean, for example, that buffalo and pigs were generally rarer near LSA sites but that more effective LSA hunting overcame the environmental difference. Our point here is not to argue that MSA people were inferior hunters. That issue remains debatable. Rather, we are concerned with how to appropriately use the zooarchaeological record to * Corresponding author. E-mail address: rklein@stanford.edu (R.G. Klein).

EGINNING in the middle 1930's and at an increasing rate in recent years reports have issued from the Soviet Union concerning the discovery of Chellean and Acheulean localities there. This paper is intended to be a critical survey of these reports. In beginning research for this paper I relied heavily on N. A. Beregovaya's Paleoliticheskie mestonakhozhdeniya SSSR (Paleolithic localities of the USSR) (1960) for pertinent bibliography. Later I supplemented Beregovaya bv examining those Soviet serial-publications which regularly contain information on the local Paleolithic (see the list of these under Abbreviations on p. 38). Every major newly discovered Paleolithic occurrence is soon reported in one or more of these serials and even relatively unimportant finds are eventually at least mentioned. I paid particular attention to numbers issued since the appearance of Beregovaya's book, and, as a consequence, several so-called Chellean and Acheulean localities which were discovered and/or described only after 1960 are considered here. Although it is unlikely that I am aware of (much less have examined) every Russian source with data pertinent to the topic of this paper, I believe I have inspected all the most important ones and certainly a sufficient number to make the conclusions presented in the final section highly probable. The remainder of this paper consists of two principal parts. The first considers reported occurrences of Chellean and Acheulean in the Soviet Union in some detail; the second presents a summary and conclusions. Before closing this introductory section, however, I would like to deal briefly with some problems of translation. Soviet prehistorians commonly use a number of terms for sediments and for artifacts which to translate without comment could be misleading. The sedimentological terms in question are glina, suglinok, supes', and pesok. In the body of this paper I have translated these as clay, loam, sandy loam, and sand, respectively (translations suggested by prominent Russian-English dictionaries). It must be noted, however, that glina and pesok do not necessarily mean deposits of purely clay-size or sand-size particles; also labeled as glina and pesok are deposits where small admixtures of coarser or finer particles (as the case may be) are present and recognizable. And suglinok and supes' are not restricted to being textural terms for soils (as loam and sandy loam generally are in the United States); rather they are used to refer to the

Twenty-one years ago, a landmark exploration of mitochondrial DNA diversity popularized the idea of a recent African origin for all living humans.1 The ancestral African population was estimated to have existed 200 ka (thousands of years ago) plus or minus a few tens of thousands of years. A corollary was that at some later date the fully modern African descendants of that population expanded to swamp or replace the Neanderthals and other nonmodern Eurasians. The basic concept soon became known as “Out of Africa,” after the Academy Award winning film (1985) that took its title, in turn, from Isak Dinesen's classic autobiography (1937). Many subsequent genetic analyses, including those of Ingman and coworkers2 and Underhill and coworkers,3 have reaffirmed the fundamental Out of Africa model. The fossil and archeological records also support it strongly. The fossil record implies that anatomically modern or near-modern humans were present in Africa by 150 ka; the fossil and archeological records together indicate that modern Africans expanded to Eurasia beginning about 50 ka.

Two competing hypotheses have long dominated specialist thinking on modern human origins. The first posits that modern people emerged in a limited area and spread from there to replace archaic people elsewhere. Proponents of this view currently favor Africa as the modern human birthplace.1–5 The second suggests that the evolution of modern humans was not geographically restricted, but invlved substantial continuity between archaic and modern populations in all major regions of the occupied world.6–7 Based solely on the fossil record, both hypotheses are equally defensible, but the spread-and-replationships scenario is far more strongly supported by burgeoning data on the genetic relationships and diversity of living humans.8–16 These data impy that there was a common ancestor for all living humans in Africa between 280,000 and 140,000 year ago, and that Neanderthals and other archaic humans who inhabited Eurasia during the same interval contributed few, if any, genes to living peiple. I argue here that the spread-and-replacement hypothesis is also more compatible with a third line of evidence: the spread-and-replacement hypothesis is also more compatible with a third line of evidence: the archeological record for human behavioral evolution.

Together with human fossils from eastern and northern Africa, southern African specimens show that anatomically modern or near-modern people were present by 100,000 years ago, when only the Neandertals occupied Europe and different, equally nonmodern people lived in eastern Asia. However, the artifacts found with early modern or near-modern African fossils imply nonmodern, Neandertal-like behavior. Artifactual markers of fully modern behavior appeared in Africa between 50,000 and 40,000 years ago, and only then were modern Africans able to expand to Eurasia, where they swamped or replaced the Neandertals and other nonmodern humans. Archaeological food debris from the western and southern coasts of South Africa suggest that an enhanced ability to hunt and gather accompanied the artifactual advance after 50,000 years ago.

Human paleontology and Paleolithic archeology both focus on the most ancient phases of human prehistory, and they often draw their materials from the same sites. Yet, since the mid-19th century when both fields were founded, their practice and theory have remained largely separate. As the 21st century dawns, human paleontologists are trained mainly in human anatomy or biological anthropology, and they take their theory principally from late 20th century evolutionary biology. In contrast, Paleolithic archeologists are trained to recover and sort ancient material residues, and their interpretative frameworks stem primarily from cultural anthropology. The difference in practice means that human paleontologists are usually ill-equipped to assess controversies in Paleolithic archeology and vice versa. The difference in theory is more profound, since human paleontologists unflinchingly attribute major morphological changes or differences to natural selection, mutation, gene drift, or gene flow. Paleolithic archeologists in contrast tend to ascribe major behavioral changes to newly developed cultural strategies or to population growth, even when the changes coincide with conspicuous morphological changes. Many especially reject biological explanations for perceived behavioral differences between modern humans and their immediate precursors, and some believe that such explanations smack of racism.1 Yet those who accept the reality of human evolution must also accept that some ancient humans must have lacked the inborn capacity to participate in historic human societies. Since we lack historic analogs for such premodern people and archeology cannot substitute for an eyewitness account, we can

A fragmentary temporal bone and partial atlas from the Middle Stone Age (MSA) at Klasies River Mouth (KRM) are described and analyzed. The atlas (SAM-AP 6268) is comparable to Levantine “Early Modern”, Neandertal and recent human vertebrae. The temporal (SAM-AP 6269) is similar to recent African homologues except that the posteromedial wall of the glenoid fossa is composed entirely of the squamous temporal, a situation that appears to be infrequent among other Pleistocene fossils. The KRM glenoid fossa is also mediolateraly broad and anteroposteriorly short in comparison with many, but not all recent specimens. Nevertheless, the KRM temporal is decidedly modern, both morphologically and metrically, by comparison with other Pleistocene specimens. The limited evidence provided by this bone is consistent with that of other MSA cranial remains from this site in suggesting an overall, if somewhat ambiguous pattern of morphological modernity.

The faunal remains from the Middle Stone Age (MSA) layers at Klasies River Mouth (KRM) suggest that South African MSA people hunted and gathered less effectively than their Later Stone Age (LSA) successors. The newly enlarged faunal sample from the MSA layers of Die Kelders Cave 1 (DK1) supports the same basic conclusion. In particular, unlike LSA people living in comparable circumstances, the MSA inhabitants of both DK1 and KRM took many fewer buffaloes than eland, even though buffaloes were probably more common near the sites. The reason for the difference may be that MSA people lacked projectile weapons, and were thus forced to rely on relatively docile species like eland. Eland may have been particularly favored because they are very amenable to driving, and the abundance of prime-age adult individuals at both DK1 and KRM may reflect the relative ease with which eland could be driven over nearby cliffs. Also in contrast to LSA people, the MSA occupants of DK1, and especially KRM, took more subadult and adult seals compared with 9–11 month olds. In this respect, MSA people resembled the brown hyenas who accumulated seal bones in a fossil den at Boegoeberg 1 (BOG1). The relative lack of young seal bones at BOG1 probably reflects the hyenas' coastal presence throughout the year, including times when young seal wash-ups are unusual. By extension, the relative rarity of young seals in MSA ( vs. LSA) sites may mean that MSA people failed to focus their coastal visits on the season of peak young seal availability. Together with artefactual and other faunal contrasts between the MSA and LSA, the differences in large bovid and seal exploitation help explain why MSA people did not spread from Africa, even though they were anatomically more modern than their Eurasian contemporaries.

The fossil record suggests that modern human morphology evolved in Africa between 150,000 and 50,000 years ago, when the sole inhabitants of Eurasia were the Neanderthals and other equally nonmodern people. However, the earliest modern or near-modern Africans were behaviorally (archaeologically) indistinguishable from their nonmodern, Eurasian contemporaries, and it was only around 50,000-40,000 years ago that a major behavioral difference developed. Archaeological indications of this difference include the oldest indisputable ornaments (or art broadly understood); the oldest evidence for routine use of bone, ivory, and shell to produce formal (standardized) artifacts; greatly accelerated variation in stone artifact assemblages through time and space; and hunting-gathering innovations that promoted significantly larger populations. As a complex, the novel traits imply fully modern cognitive and communicative abilities, or more succinctly, the fully modern capacity for Culture. The competitive advantage of this capacity is obvious, and preliminary dates suggest that it appeared in Africa about 50,000 years ago and then successively in western Asia, eastern Europe, and western Europe, in keeping with an African origin. Arguably, the development of modern behavior depended on a neural change broadly like those that accompanied yet earlier archaeologically detectable behavioral advances. This explanation is problematic, however, because the putative change was in brain organization, not size, and fossil skulls provide little or no secure evidence for brain structure. Other potential objections to a neural advance in Africa 50,000-40,000 years ago or to the wider “Out-of-Africa” hypothesis, include archaeological evidence (1) that some Neanderthals were actually capable of fully modern behavior and (2) that some Africans were behaviorally modern more than 90,000 years ago.

Die Kelders Cave 1, on the southwestern coast of South Africa, preserves a rich Later Stone Age (LSA) occupation and a thick series of Middle Stone Age (MSA) layers below. Limited excavation of the MSA layers has yielded numerous lithic artifacts and faunal remains, and nine human teeth. Sedimentologial and faunal evidence suggests that the MSA levels accumulated under comparatively cool, mesic conditions, probably at the beginning of the Last Glaciation (isotope stage 4). The MSA artifact assemblage consists overwhelmingly of quartzite debitage. Elongate flakes are fairly common; systematic retouch is extremely rare. The high frequency of silcrete artifacts in some of the lower units may signal the Howiesons Poort variant of the MSA, although diagnostic backed and truncated pieces are absent. The animal bones come from rodents, insectivores and other small creatures that were probably introduced mainly by owls, as well as from larger mammals and seabirds that were probably introduced mainly by humans. This latter component resembles the Klasies River Mouth MSA fauna in the abundance of eland relative to wild pigs, the dominance of penguins over flying birds, and the absence of fish. At both Die Kelders and Klasies River Mouth these features tentatively suggest that MSA hunter-foragers exploited animal resources less effectively than their LSA successors. Although the Die Kelders MSA human teeth tend to be somewhat larger, they are morphologically similar to modern African homologues, and they exhibit several features that tend to distinguish modern Africans among other populations. These teeth may be added to the small fossil sample that attests to the morphological modernity of the MSA inhabitants of southern Africa.

The Elandsfontein site, Western Cape Province, South Africa, is well known for an archaic hominin skullcap associated with later Acheulean artifacts. The site has also provided nearly 13,000 mammalian bones that can be identified to skeletal part and taxon. The assemblage derives from 49 species, 15 of which have no historic descendants. Comparisons to radiometrically dated faunas in eastern Africa indicate an age between 1 million and 600 thousand years ago. Unique features of the fauna, including the late occurrence of a dirk-toothed cat and a sivathere, may reflect its geographic origin in a region that was notable historically for its distinctive climate and high degree of biotic endemism. Together, taxonomic composition, geomorphic setting, and pollen extracted from coprolites indicate the proximity of a large marsh or pond, maintained by a higher water table. The small average size of the black-backed jackals implies relatively mild temperatures. The sum of the evidence places bone accumulation during one of the mid-Pleistocene interglacials that were longer and cooler than later ones, including the Holocene. The geomorphic context of the fauna presents no evidence for catastrophe, and most deaths probably resulted from attritional factors that disproportionately killed the young and old. However, only the dental-age profile of long-horned buffalo supports this directly. Field collection methods biased skeletal-part representation, but originally, it probably resembled the pattern in the younger, marsh-edge Acheulean occurrence at Duinefontein 2, 45 km to the south. Excavation there exposed multiple vertebral spreads, which probably mark carcasses from which hominins or large carnivores removed the meatier elements. Bone damage at both sites suggests that, despite abundant artifacts, hominins were much less important than carnivores in the bone accumulation. Together with limited observations from other sites, Elandsfontein and Duinefontein provisionally suggest that Acheulean-age hominins obtained few large mammals, whether by hunting or scavenging. 2006 Elsevier Ltd. All rights reserved.

Die Kelders Cave 1, first excavated under the direction of Franz Schweitzer in 1969-1973, was re-excavated between 1992 and 1995 by a combined team from the South African Museum, SUNY at Stony Brook, and Stanford University. These renewed excavations enlarged the artefactual and faunal samples from the inadequately sampled and less intensively excavated lower Middle Stone Age (MSA) layers, increased our understanding of the complex site formation processes within the cave, enlarged the hominid sample from the MSA deposits, and generated ESR, TL, and OSL dates for the MSA layers. Importantly, these new excavations dramatically improved our comprehension of the vertical and lateral characteristics of the MSA stratigraphy. Surface plotting of the MSA layers has led to the identification of at least two major zones of subsidence that significantly warped the layers, draping some along the eroding surface contours of major blocks of fallen limestone roof rock. A third zone of subsidence is probably present in the older excavations. Dramatic roof falls of very large limestone blocks occurred at least twice-once in the middle of Layer 4/5 where the roof blocks were only slightly weathered after collapse, and at the top of Layer 6 where the blocks weathered heavily after collapse, producing a zone of decomposed rock around the blocks. Many of the sandy strata are cut by small and localized faults and slippages. All of the strata documented by Schweitzer's excavations are present throughout the exposed area to the west of his excavated area, where many of them thicken and become more complex. Layer 6, the thickest MSA layer, becomes less diagenetically altered and compressed to the west.

Die Kelders Cave 1, South Africa, has provided more than 150,000 taxonomically identifiable mammal and tortoise bones from Middle Stone Age (MSA) and Later Stone Age (LSA) deposits. Cape dune mole-rats dominate the mammal sample, and they appear to have been accumulated mainly by people during the LSA occupation and mainly by eagle owls in the MSA. In sharp contrast to the LSA fauna, the MSA sample contains extralimital ungulates that imply relatively moist, grassy conditions. The large mean size of the MSA mole-rats also points to greater humidity, while the large size of the gray mongooses implies cooler temperatures. The sum supports luminescence and ESR dates that place the MSA occupation within the early part of the Last Glaciation (global isotope stage 4). The Die Kelders ungulate bones support those from Klasies River Mouth in suggesting that MSA people obtained dangerous terrestrial prey much less frequently than their LSA successors, probably because MSA people lacked the bow and arrow and other projectile weapons. The Die Kelders tortoise bones constrain the extent of climatic change, since their abundance indicates that warm, dry days remained common, at least seasonally. The tortoises tend to be much larger in the MSA layers than in the LSA ones, suggesting that MSA people collected tortoises less intensively, probably because MSA populations were relatively sparse.

Excavations at Duinefontein (DFT) 2 near Cape Town, South Africa have recovered numerous stone artefacts and animal bones on an ancient surface sealed within iron-stained eolian sands. U-series analysis of an overlying calcrete places the sands before 150 ka ago, while the large mammal taxa imply an age between 400 and 200 ka ago. The artefacts include a classic Acheulean handaxe and probable biface shaping flakes that support this age estimate. The principal mammalian species are long-horned buffalo, black wildebeest, greater kudu, Cape zebra, and grysbok/steenbok, which imply a grass-and-bush mosaic instead of the historic small-leafed shrubland. Hippopotamus and reedbuck indicate that water stood nearby, probably in dune swales. The large mammal bones are mostly vertebrae and other axial elements, often in near-anatomical order. Both proximal and distal appendicular elements are rare. Bones with carnivore damage are common, but ones with stone tool marks are scarce. The sum suggests a water-edge attritional death site where people played a minimal role and carcasses were disarticulated mainly by carnivore feeding and by trampling. Stone tool marks tend to be equally rare at other Acheulean attritional death sites, and the implication may be that Acheulean people rarely obtained large mammals, whether by hunting or scavenging. Human scavengers at DFT2 would not have encountered a disproportionate number of distal (versus proximal) limb elements, and it follows that the tendency for distal elements to dominate many archeological assemblages need not reflect scavenging versus hunting. Even if DFT2 was not itself a locus of intense human activity, it provides a useful baseline for evaluating bone damage, skeletal part representation, and other variables at sites where people were deeply involved.

Human origins research by archaeologists has expanded the evidence of the diet and subsistence activities of ancient hominids. We examine an important component of that evidence, the 1.75-million-year-old faunal assemblage from the FLK Zinjanthropus site at Olduvai Gorge, Tanzania. Skeletal-part frequencies are used to evaluate hominid access to and differential transport of carcass portions of differing nutritional value. Cut-mark frequencies and locations are used to evaluate butchery patterns including skinning, disarticulation, and defleshing of carcasses. In contrast to other recently published assessments of the FLK Zinjanthropus data, we conclude that (1) ancient hominids had full access to meaty carcasses of many small and large animals prior to any substantial loss of meat or marrow bones through other predator or scavenger feeding; (2) ancient hominids were butchering animal carcasses by an efficient and systematic technique that involved skinning, disarticulation, and defleshing; and (3) the FL...

This paper discusses prehistoric man‐environment relationships in the Southern Cape, defined as Africa south of 32°S. lat. Although the area was occupied in the early Pleistocene and perhaps earlier, archaeological evidence only becomes substantial in mid‐Pleistocene contexts. Local mid‐Pleistocene peoples (Acheulean) seem to have been widespread, but reconstruction of their environment and subsistence remains a goal of future research. Detailed archeological knowledge is so far restricted to the Upper Pleistocene and Holocene, but even here major questions go unanswered. Chief among these is the nature of the replacement of the Middle Stone Age (MSA), some 30–40,000 years ago. In all known Southern Cape sites, the MSA is followed by a gap in occupation, varying from 15,000 to 25,000 years. This gap alone suggests that the immediate successors to MSA peoples may have been characterized by significantly different settlements and subsistence patterns from those of their predecessors. Additionally, ...