Early hominids began leaving Africa for Asia almost one million years earlier than previously thought. What drove these migrations?

This article was published by American Scientist, November-December 1996.

      More than a century ago, Dutch paleontologist Eugene Dubois suggested that human origins lay in Southeast Asia, and he soon found the undeniably earliest hominid skeletal remains on the island of Java. In the 1930s, many more fossils of similar primitive character came to light near Beijing, and the entire Asian collection became known as Homo erectus. Presumably arising from an Asian ape, "upright man" had evidently occupied a great swath of eastern Asia, and provided the logical precursor to the more advanced and younger Neanderthal and Cro-Magnon (Homo sapiens) fossils of Europe. In evolutionary terms, Homo erectus was thought to have emerged in Asia and later dispersed to Europe.
      During the 1960s, the eastern Rift Valley region of eastern Africa began to yield contrary evidence in its many much older Australopithecus fossils as well as an ample number of fossils from Homo. By the 1970s, human origins were believed to lie in Africa, with a much later dispersal toward East Asia. Movement from Africa to Europe came yet later. The new African research, a collaborative effort among paleontologists, geologists, paleoclimatologists and others, also began to ask about the timing and cause of human emergence. In the light of new ecological theory and increasingly sensitive environmental evidence, paleoanthropologists could now advance relatively complex behavioral models.
      Current evidence suggests that about 3.0 to 2.4 million years ago (mya), the relatively cool, dry climate of tropical Africa presented challenging new conditions for woodland-dwelling Australopithecus. In theory, the cooler climate cleared out some of the woodland to form new open habitats. It is at about this time that the earliest species of the genus Homo emerged, between 2.5 and 2.0 mya, to exploit the new habitats as a rather aggressive omnivorous scavenger.
      Climate has been less useful for understanding the intercontinental dispersion of Homo. Until recently, the earliest Homo fossils in Asia appeared to be no more than 1.1 million years old, representing a time well after emergence and not directly related to significant climatic events. Paleoanthropologists have therefore explained dispersion as a separate stage of development, and as the result of "internal" factors, such as population saturation and technological advances in tool making and resource scavenging.
      Recently, however, we and our Chinese colleagues have contributed to what has become a wave of new Asian fossil discoveries and technical re-analyses that change this picture. In tropical and subtropical East Asia, the age of newly discovered fossils of Homo and simple stone tools, as well as some revised dates for known remains now approach 2 million years, nearly 1 million years older than previous estimations. In temperate west Asia, a new hominid and associated tools reach 1.4 million years in age. This new evidence extends the pattern of well-known contemporary eastern Mediterranean archaeological sites northward to suggest a later--but still quite early--movement to the more temperate areas of the Middle East and mid-latitude west Asia.
      The new finds and age determinations give distinctness and complexity as well as antiquity to the formerly late and amorphous pattern for Asian dispersal. Our interest lies in the initial dispersal of early Homo from tropical Africa eastward across tropical and subtropical habitats of south and East Asia. In fact, early Homo now seems to have arrived in East Asia so early that its African emergence and initial subtropical Asian dispersal must be linked. Thus the factors that triggered the evolution of Homo from Australopithecus also encouraged early Homo to leave Africa, at least initially. We reconsider the new evidence for early dispersals in the light of climatic, morphological, technological and behavioral factors hypothesized for the emergence of Homo. We find that the striding gait, the elementary stone tools and the simple, but expansive, pattern of scavenging that characterizes the emergence of Homo also served its initial dispersal. In contrast, later populations of Homo colonized more temperate habitats under more complex and less obvious conditions.

An early hominid, Homo ergaster, depicted in this diorama from the American Museum of Natural History's Hall of Human Biology and Evolution, lived nearly 2 million years ago in the eastern Rift Valley of Africa. Until recently, anthropologists thought that such early hominids did not disperse from Africa to Asia until 1 million years ago. New fossil finds and dates from Asia, including the authors', now suggest that early Homo arrived in East and South Asia by 2 million years ago. Tying the new evidence to paleoclimate and ecological theory, the authors suggest that the physical adaptations of African emergence-among them, a ranging bipedal gait, stone technology, increased intelligence and extensive scavenging-may also have enabled early Homo to colonize subtropical Asia very quickly.

Time, Climate and Species

      The most basic questions for human dispersal have remained hypothetical during the past two decades. When did hominids first leave Africa? Which species was the first to leave? Why did they leave? The issue of age has always overshadowed all others. In the eastern Rift Valley sites, fossils are usually recovered from relatively fine-grained deposits laid down by water and wind. These formations often include layers of volcanic ash that are easily dated using the potassium-argon (K-Ar) radiometric method. Alternatively, in Europe and subtropical Asia many fossils are found within the diverse and complex deposits that accumulate in caves, where depositional histories are difficult to interpret and volcanic materials are not present. The net effect is that the age of Homo fossils has been measured more precisely (and consistently older) in Africa than in Eurasia. Thus, the earliest Homo erectus (better termed Homo ergaster) fossils in the eastern Rift Valley appear fully developed by 1.9 mya, whereas the Javanese fossils (classic Asian Homo erectus), which are thought to be the earliest in Asia, have traditionally received broad age estimates of only 700,000 years to 1.1 million years. The nearly one-million-year disparity between the African emergence and the initial Asian arrival has for years been the basis of the conventional theory for a late dispersal.
      Recent developments in techniques that provide the absolute ages of artifacts--such as paleomagnetism, electron-spin resonance (ESR), and single-crystal argon (Ar/Ar) methods--have shed new light on the arrival of Homo in Asia. Moreover, the discovery of new artifact-bearing sites makes the dispersal of Homo a much more accessible question. At Riwat and Pabbi Hills in Pakistan, simple stone tools have a paleomagnetic age of about 1.9 million years. At Sangiran and Mojokerto in Java (Indonesia) the sedimentary contexts for three well-known cranial specimens of Homo erectus now have Ar/Ar age determinations of 1.6 to 1.8 million years. The most intriguing of the new finds for early Asian dispersal come from Longgupo, a cave in southeastern Sichuan Province, China. The Longgupo hominid teeth have affinities to early African Homo and the stone artifacts resemble early African tools. Last year, we and our Chinese colleagues, Huang Wanpo and Gu Yumin at the Institute for Vertebrate Paleontology and Paleoanthropology, published ESR and paleomagnetic analyses that indicated an age of 1.9 mya for the Longgupo remains. The growing number of Asian hominid fossils and stone-tool assemblages that approach 2.0 million years in age now suggests that an early population of Homo arrived in eastern Asia within a few hundred thousand years of arising in Africa. In the light of this new evidence for early dispersals, it appears Homo emerged not so much in adaptation to challenging conditions, but fully poised to dominate new resources in new territories.

Figure 2. Evolution of the genus Homo from Australopithecus seems to be linked with global climate changes between 2 and 3 million years ago. Cooler temperatures diminished the tropical-woodland habitat of Australopithecus in favor of more open savanna. Several species of Homo evolved rapidly to occupy these new habitats, as did Paranthropus , another descendant of Australopithecus . Paranthropus became extinct by 1.2 million years ago. Early Homo maintained a generalized anatomy as it spread throughout tropical and subtropical Asia, but apparently became specialized with the evolution of Homo erectus by 1.8 million years ago. The much later European dispersal of African Homo heidelbergensis also seems to have resulted in the specialized Neanderthals. All of these early dispersals of Homo were apparently eclipsed by the late Eurasian dispersal of Homo sapiens , from Africa, some 100,000 years ago. The three darker bands indicate periods of greater relative aridity.

Figure 3. Hominids now known as Homo erectus were found on Java, Indonesia, in 1891, and at Zhoudoudian, near Beijing, in the 1930s. As Homo erectus was clearly more primitive than hominid fossils known in Europe, human beings were initially thought to have emerged in East Asia and dispersed westward. Since the early 1960s, numerous fossils from African localities in the eastern Rift Valley, Lake Malawi and South Africa have demonstrated an African emergence for Homo . In the 1990s, advances in dating methods and new finds at Dmanisi (Georgia), Riwat (Pakistan), two Javanese sites and Longgupo (China) show that early Homo had arrived in East Asia by 2 million years ago. Areas in pale blue indicate land masses submerged since those early dispersions.

      Crucial aspects of hominid evolution and dispersal evidently relate to global climatic trends, particularly to the cooling and drying associated with glaciation in the northern hemisphere. Within the Rift Valley, such oscillations repeatedly added open components to woodlands, to effect more mixed or mosaic-like habitats. By implication, climatic trends prompted the large hominoids to develop new physical and behavioral adaptations, but the process and results are not always clearly related. For example, a cooling event during the Late Miocene epoch (about 6.0-5.3 mya) correlates well with the time at which apes and hominids diverged as estimated by molecular clocks. Although climate must have played a role in this speciation event, fossils from this period have yet to be found. Consequently, the crucial anatomical effects of climatic cooling and their relation to the divergence of the apes and the hominids are not known.
      Fortunately, it is easier to link the emergence and dispersal of Homo with cooling during the Middle Pliocene (3.0-2.4 mya). During this period in Africa, many mammalian species were pushed toward extinction, speciation or dispersion. Yale University paleontologist Elisabeth Vrba has documented extinctions for larger species of forest-adapted African bovids, particularly the antelopes, and the emergence of more cursorial, open-dwelling species that occupy grassland habitats to this day. Vrba also has shown that some six species of African bovids, a relatively large number for any period, dispersed to Eurasia during the Middle Pliocene. The dispersal of large bovids is especially interesting because their shift from the forest to the open resembles that hypothesized for the hominids. Moreover, since open-country bovids could have been hunted or scavenged by Homo, the two groups may have emerged and dispersed together.
      Regarding the hominids, Middle Pliocene cooling underlies the divergence of Australopithecus into two evolutionary lines. One yielded Paranthropus, whose robust jaws and massive teeth reflect a rather specialized coarse vegetarian diet. In this robust line, Paranthropus boisei survived until 1.2 mya and may have developed stone tools. Nevertheless, in other ways Paranthropus was not substantially different from its australopith ancestors. With sexually dimorphic bodies and ape-sized brains, Paranthropus was confined to tropical habitats and never encountered its continental limits.
      The second, more omnivorous line issued Homo in the African fossil record. From the discovery of Javanese finds in 1891 to the early 1960s, when Louis and Mary Leakey's work began to pay off in spectacular finds at Olduvai Gorge, our own genus comprised just two species: Homo erectus (identified only later in Africa and, by some accounts, in Europe) and our own Homo sapiens. With the Leakeys' discoveries of Homo habilis at Olduvai, Homo erectus's direct antecedent seemed to have appeared at 1.8 to 1.6 mya. When first found, Homo habilis presented a larger cranium and narrower teeth than the older autralopiths and yet still showed some primitive features such as small size. As fossil finds of Homo habilis have accumulated more recently, apelike long arms, short legs and australopith-like thigh anatomy have cast doubt on this relationship. Homo habilis is morphologically too primitive to be an ancestor of Homo erectus.
      Concurrently, a number of fairly complete and well-preserved fossils from the Turkana Basin of northern Kenya (as well as a skull from Swartkrans, South Africa) indicate that Asian Homo erectus has an equally ancient sister species in Africa. By 1.9 mya Homo ergaster exhibited limb proportion and body size comparable to Homo sapiens and to Homo erectus, and with a cranium larger than that of Homo habilis. Comparing the African and Asian forms yields subtle but significant differences. The skull of Homo ergaster is more generalized, having a higher or domed cranium, fairly thin cranial bones, weak brow ridges and a lightly built face, features that align the species more with Homo sapiens. Alternatively, Asian Homo erectus (cover image) has always been defined on rather specialized features including a long cranium, a low forehead, thick cranial bones, large projecting brow ridges and a heavier face (compared to Homo sapiens). With the new earliest dates for Homo erectus in Asia at 1.8 mya, the two species seem to be evolutionary contemporaries (many do not even term them separate species). Although it is tempting to consider Homo ergaster as the more generalized African form of Homo erectus, and therefore the first colonizer of Asia, it is more likely that an older species ancestral to both forms left Africa more than 2.0 mya.
      Scant evidence for this pre-erectus hominid--an emergent "early" Homo--come from four other areas of the eastern Rift in addition to the Turkana Basin: the Hadar Basin of northeastern Ethiopia, the Omo Valley of southwestern Ethiopia, the Baringo locality of central Kenya and the Uraha locality of eastern Malawi. These areas have geological formations dating to middle-late Pliocene (2.5 to 1.8 mya), the end of the critical Pliocene cooling and the beginning of a period of climatic stability. Some early fossils have been termed Homo rudolfensis, a taxon having affinities with Homo habilis and possibly with Homo ergaster/erectus. Other finds (for example, the isolated teeth from Omo and the partial temporal bone from Baringo) are too fragmentary to classify with any specificity. In Hadar and in Omo, early Homo is associated stratigraphically with emergent stone-tool assemblages. Although a detailed understanding of early Homo awaits more fossil discoveries, the significant date of origin and the climatic link are incontrovertible.

Please click on an image above to see the full size version..

      As typified by Homo ergaster, early Homo was the first hominid to develop a ranging bipedal gait, one that allowed it to cover a lot of ground in a short period. Long, well-muscled limbs on a lanky torso also conferred physical strength and defensive presence. Leverage and strength in the arms and hands gave early Homo the ability to make simple chopping and cutting implements. Whereas Australopithecus and Paranthropus had significant sexual dimorphism in body size and strength, early Homo did not. Among other primates, reduced sexual dimorphism usually means that both sexes perform similar economic activities, that males compete less physically among themselves for females and that males and females pair-bond for long periods. An enlarged cranium also typified early Homo. A larger brain may have bestowed a more flexible intelligence, useful for finding resources within new habitats, as well as the complex behavior Homo came to use against prey and predators.
      The early specimens of Homo also exhibit the relatively diminished premolars and molars-the cheek teeth-of an omnivorous eater, one for which animal protein played a significant dietary role. As a part-time meat eater, early Homo probably relied on large carnivores to supply many usable packages of animal protein-very pragmatically scavenging the remains of what real carnivores could more effectively hunt. At the other dietary extreme, these hominids certainly consumed some hard, tough plant foods. In contrast to Paranthropus, whose digestion of such foods probably began inside its mouth, early Homo more often and more completely processed difficult animal and vegetal resources with stone-tool technology, which was employed to break, crush, split and cut up hard foods before ingesting them. Rather than being implements of predation, early tools underlay a technology that essentially added a new first stage to digestion. Indeed, we hypothesize a close relationship between Pliocene hominid biology and a very elementary technology, in effect a "biological technology" in which both the mass and the jagged edges of a few chipped rocks gave early Homo access to a wide range of nutritional resources. Simple stone tools represent immediate extensions of the forelimb and hand for breaking down or processing tough foodstuffs.

Earliest complete skeleton of Homo is that of a boy from the West Turkana region of Kenya. Homo ergaster already had many anotomical features of Homo sapiens, including a high, domed cranium with realatively light cranial and facial bones and a lanky torso with long, well-muscled limbs. The overall size and proportions of its skeleton gave Homo ergaster a ranging, bipedal gait as well as physical strength and a defensive presence. Leverage and strength in the arms and hands helped this hominid use simple stone tools effectively to chop and crush food. (photograph provided by Alan Walker at the Pennsylvania State University and reproduced with permission from the National Museums of Kenya.)

   

  Hominid biological remains may be difficult to classify, but stone tools present even greater problems of interpretation. One way to understand the development of early stone technology is to envision emergent and advanced stages separated by a transition. Currently, evidence for the emergent stage appears with the late-middle Pliocene (about 2.5 mya) in Hadar, Omo and Turkana. Mzalendo Kibunjia of the National Museums of Kenya has proposed the name Omo Industrial Complex for such assemblages, in which rocks were broken or casually chipped into very basic implements. Only a few general tool types can be defined for the Omo-type assemblages: simple core choppers and rough flake scrapers. The basic technological characteristics of Omo-type core-flake assemblages vary by the raw materials available in each region. In Hadar and Turkana, where large volcanic cobbles are present, the tools tend to be large and the chipping technique a little more complex. In Omo, where small, tough quartz cobbles were used, the tools are much smaller and more haphazardly made. At present, the Omo-type localities and tools represent the initial threshold of stone technology at a date that correlates well with the emergence of Homo itself.
      The advanced or Acheulian stage begins with the early Pleistocene, about 1.6 to 1.4 mya-well after Homo ergaster attains full development in the Turkana Basin. The Acheulian technological complex is achieved as the selection of raw materials, the preparation of the stone core and the chipping procedures become much more complex, and the tools themselves become somewhat specialized. Bifacial chipping is the hallmark of the Acheulian Industry. With this technique, a tool blank is chipped from two directions across a bisecting plane. The blank is worked around a circumferential edge to resemble a plump discus, or a double-sided tortoise shell, and the entire edge becomes the working part. Although earlier hominids had developed crude bifacial techniques by 2.0 mya, the method did not become the basis for a distinctive set of Acheulian biface tools until about 1.5 mya. The Acheulian appears to emerge in the eastern Rift in areas such as Konso-Gardula in Ethiopia, as well as Peninj and Olduvai Gorge in Tanzania. Nevertheless, Acheulian bifaces are found as far north as the Jordan Valley of Israel (at 'Ubeidiya) by 1.4 mya. When hominid remains and early Acheulian tools are associated within any site stratum in Africa, the species is always Homo ergaster or Homo erectus, not Homo habilis or Paranthropus boisei.
      The period from about 2.0 to 1.5 mya is best seen as a long and important transition between the advent of chipping techniques and the achievement of standardized Acheulian biface tools. Much of this period is represented in the lower beds of Olduvai Gorge, Tanzania, where Mary Leakey of the National Museums of Kenya defined the well-known Oldowan Industry more than 30 years ago. At the time, the Oldowan was thought to be the earliest manifestation of stone technology and, indeed, the earliest assemblages in Bed I (the oldest levels of Olduvai) resemble those from Ethiopia and northern Kenya now deemed oldest. Nevertheless, most Oldowan tool kits reflect more care and skill in choosing raw materials and in preparing and striking cores to create usable flakes and core tools than do the Omo-type assemblages. Over the several hundred thousand years evident in Olduvai's stratigraphy, Oldowan assemblages undergo distinct refinement in chipping techniques and some standardization in tool form. By 1.7 to 1.6 mya, bifacial tools help to define the Developed Oldowan Industry. At Olduvai, the initial stone tool finds came in stratigraphic association with a new hominid, one with a larger brain and more gracile features than had the previous robust australopith skull found at Olduvai Gorge (Paranthropus boisei). The clear association of an advanced hominid with stone tools prompted the Leakeys to designate the new species Homo habilis, literally "handy man." It becomes clear that more than one species of Homo--and probably Paranthropus boisei--made Oldowan-type tools at Olduvai Gorge and elsewhere.
      Given its relatively early appearance and greater complexity over the earlier stages as well as its association with Homo ergaster, the African Acheulian technology has often played into hypotheses for Eurasian dispersal. However, new dates and technological analyses make any such role unlikely in Asia. Although Acheulian tools always seemed to antedate the earliest tool assemblages in East Asia, the recently discovered stone tools at Longgupo and Riwat are as ancient and as simple in design as their Omo and early Oldowan counterparts in Africa. None of the older Asian assemblages contain handaxes, and few exhibit even the standardized chipping patterns of the Developed Oldowan or the Acheulian technologies.
      The current and revolutionary evidence for a very early dispersal of hominids from Africa to Asia may be reiterated as follows. Fragmentary fossils representing the emergent genus Homo are consistently dated to nearly 2.5 mya at various points in the eastern Rift Valley. Likewise, as the earliest stone tools, also found in the eastern Rift, have equal antiquity, the emergence of one must be linked to the other. By 1.9 mya, Homo ergaster presents undeniable morphological features for moving great distances: long torso and limbs, narrow hips, a large brain and reduced dentition. With the new evidence from Longgupo, Java and Riwat, it becomes clear that early Homo (the immediate ancestor to Homo ergaster and Homo erectus) and simple stone tools arrived in tropical and subtropical Asia by about 2.0 mya. The emergence of Acheulian technology in east Africa confirms the hypothesis for early Asian dispersal. Distinctive bifaces date to 1.5 mya in the eastern Rift Valley and the Middle East, and to 600,000 years ago in Europe. The absence of Acheulian bifaces at the early sites in south or East Asia suggest that Homo must have initially left Africa before the Acheulian stage appeared in Africa. Consequently, early Homo dispersed to subtropical Asia with a very elementary technology.

 

Stone technology appears to have emerged in the eastern Rift Valley about 2.6 million years ago. At this time, tools of Omo Industrial Complex appeared, in which early Homo toolmakers chose a durable crystalline stone, often picked off the ground as a fist-sized cobblestone, to break or chip into a core and several flakes (top).

By 1.6 to 1.4 million years ago, raw-material selection, core preparation and chipping techniques had improved markedly. In reorienting the core and hammerstone, Homo ergaster toolmakers could chip flakes from a lenticular cobble around its periphery on both faces (bottom ). Competent bifacial chipping, the hallmark of Acheulian artifacts, was a skill carried to Europe by Homo heidelbergensis about 600,000 years ago. Having attributes of both Omo and Acheulian tools, Oldowan artifacts, first recovered from Olduvai Gorge in the 1960s, reflect an intermediate technological stage lasting from about 2.0 to 1.5 million years ago.

Figure 7. Early stone artifacts show technological development in Africa and the Middle East, but none in subtropical Asia, where stone technology seems to have arrived and remained at Omo and Oldowan levels. Hammerstone and core tool from Olduvai Gorge, Tanzania (a ) helped define the style of tool called Oldowan. Flake and core tools from Lokalelei, Kenya, (b) are of the Omo type. Core tools of the Oldowan type were found in 'Ubeidiya, Israel (c ), as were Acheulian bifaces (d ). Core tools were found in Longgupo, China (e ), Riwat, Pakistan (f) and Dmanisi, Georgia (g ). Bars connect two views of the same tool. Scale is 1 inch = 15 centimeters

      In scavenging increasingly open terrain, early Homo was subject to two critical environmental factors. One was access to basic natural resources, including water, animal carcasses and stone for tools. The other was the ability to gain refuge from the heat of the sun, the cold of the night and the threat of carnivores. Two sequential patterns of subsistence behavior arose in relation to these factors. The first characterizes the earliest stone-tool sites that appear about 2.5 mya. The early sites are found on or near important natural features such as lake margins, stream confluences or rock outcrops. We term these locales and their immediate surroundings "catchments" to denote the presence of numerous complementary resources. Catchment sites often have rather large, diverse accumulations of bone and stone refuse that indicate significant hominid occupation. It appears that early hominids congregated within catchments to use a number of resources for a fair amount of time. After exhausting one catchment's resources, the group would move on to another in a sequence of similar discrete movements. We term this form of subsistence behavior "catchment scavenging." It represents a way to use the resources of a large territory with a minimum of knowledge.
      The second subsistence pattern appears after 2.0 mya and becomes common after 1.6 mya, and is generally associated with the Acheulian Industry. These later sites are found more often at locales offering just one or two resources, and they exhibit smaller and less-generalized stone and bone assemblages. Sites of this type may encompass many such small locales across a large area. A more complex behavior best fits this evidence, one in which small groups of hominids can manage the range of resources across a larger territory. A primary component of "territory scavenging" is the ability to work among small resource locales simultaneously or interdependently instead of sequentially. Even with this advance, the intensive localized knowledge required for territory scavenging could still be amassed at a prelinguistic, presymbolic and largely unconscious level-which would characterize the hominid groups that employed this type of scavenging style.
      The distinction between catchment scavenging and territory scavenging provides a behavioral key for understanding early Asian dispersal. In leaving Africa before 2.0 mya, early Homo most likely employed the catchment strategy. In moving across numerous dispersed catchments, early Homo essentially could not invest in or manage a home territory in the way that modern Homo sapiens do. Moreover, as a spatially extensive strategy, catchment scavenging inherently dispersed hominid groups over a limited number of ideal points per given area. With the Middle Pliocene retreat of wooded refuge areas, the number of viable catchments probably diminished through time. Ranging ever more widely among fewer catchments, some hominid groups may have been forced north and east, and out of Africa before 2.0 mya. After this date, territory scavenging may have effectively kept remaining populations within the continent and raised intergroup competition for increasingly scarce resources. Territory scavenging may lie behind other particularly African technological advances, such as the development of the Acheulian Industry about 1.6 mya and the emergence and Eurasian dispersal of more advanced hominids, including Homo heidelbergensis about 600,000 years ago and anatomically modern Homo sapiens about 200,000 to 100,000 years ago.

Collapsed cave of Longgupo in eastern Sichuan Province, China, has recently yielded important hominid fossils and stone artifacts for which the authors have established a date of nearly 2 million years. The hominid remains and artifacts lie in geological association with many other mammalian fossils that provide sensitive evidence for species movement. In addition to the arrival of early Homo, the Longgupo fauna indicates that subtopical species of horses and pigs also moved to new northerly limits during the late Pliocene.

Jaw fragment from Longgupo cave (left), which dates to about 1.9 million years ago, resembles an analogous specimen (right) from East Turkana (ER 992), dated to about 1.2 million years. Both fossils show more generalized hominid dental features in molar cusp patterns and premolar crown shape when compared with Asian Homo erectus. The Longgupo jaw is thus more primitive than Homo erectus and could represent the ancestral condition for that species. The Longgupo hominid best represents a population of early Homo arriving to East Asia about the time Homo ergaster appears in Africa. (Photograph of African hominid jaw provided by Alan Walker and reproduced with permission from the National Museums of Kenya. Longgupo photograph courtsy of the authors.)

      In introducing our argument, we noted that Eugene Dubois began collecting fossil hominids on the island of Java before the turn of our own century. Since Dubois's time, more than 80 cranial specimens of have been found at six principal localities. Although the localities generally have complex stratigraphies, the fossils have been collected without reliable stratigraphic information. However, two Javanese localities, Mojokerto and Sangiran, stand out for having Homo erectus fossils associated with the Pucangan Formation, which has been dated independently at nearly 2.0 mya. Recently, Carl Swisher at the Berkeley Geochronology Center and his colleagues have reconfirmed the geological associations and have applied Ar/Ar techniques to Pucangan deposits at these sites. Sediments from the Mojokerto site, in which a well-preserved juvenile skull has been found, yield an average age of 1.81 mya. Sediment from Sangiran, in which two other crania were found, gives 1.66 mya. The new ages for these three specimens show that the low cranial vault, thick cranial bones and massive face that characterize Homo erectus evolved rapidly after early Homo arrived in Asia. Very recently, Truman Simanjuntak of the National Research Centre of Archaeology in Jakarta and François Sémah at the Musée National d'Histoire Naturelle at Paris have found simple but indisputable flake tools in the Kabuh Formation, which overlies the Pucangan. Although these artifacts were deposited more recently than the remains of the earliest hominids, the discovery puts tools in the hands of Javanese Homo erectus in general, for which any such association had been lacking.
      The technological component of early dispersal also appears southeast of Rawalpindi, Pakistan, among some large tributaries of the Indus River. Two areas, Riwat and the Pabbi Hills, have been studied by British and Pakistani specialists. A small number of simple flaked quartzite tools were found at Riwat during the 1980s, and a larger core-flake assemblage was recovered from the Pabbi Hills locality. Geological associations and a paleomagnetic analysis suggest an age of nearly 2.0 million years for the Riwat deposit, and more circumstantial evidence points to a similar age at Pabbi Hills. The consistent technological feature of all the earliest Asian stone-tool assemblages is simplicity. Indeed, all assemblages, including the later tools from Sangiran, recall stone artifact assemblages from either the Omo or Oldowan technological complexes but not from the Acheulian.
      How could groups of scavenging hominids employing the simplest of technologies spread across south and East Asia so quickly? Middle Pliocene cooling and drying also encouraged South Asian dispersal as a band of open tropical and subtropical environments appeared from the area that is now Saudi Arabia eastward south of the Himalayas and into Southeast Asia. Like the African grasslands, these habitats (of more mixed vegetation types) drew catchment-scavenging hominids. Moreover, by the late Pliocene a significant portion of the earth's moisture was locked in glacial ice mass, lowering the global sea levels to expose what is now the continental shelf as expansive areas of coastal plain or to reveal isthmuses where there are now straits. Two exposed areas would have facilitated eastward dispersal. The Bab-el-Mandab (linking Africa and Asia between modern-day Djibouti and Yemen) adjoins Hadar, with its very early stone-tool assemblages. This short isthmus would have produced the best departure point for eastward travel within the tropical zone. The presence of similar modern flora and fauna (including Hamadryas baboons) on both sides of the strait signal its importance for numerous dispersals. Departing Africa on this route, early Homo could trend east-northeast across the Arabian Peninsula to reach a much smaller Persian Gulf, possibly crossable at Hormuz. A short journey east along the Arabian Sea brings the Indus Valley and direct access to Riwat and Pabbi Hills. Thus, although these South Asian sites seem to be along way from Africa, the departure point of Hadar lies surprisingly close given late Pliocene geography. A route south of the Himalayas would have brought early Homo to Southeast Asia.
      Within the lowland Southeast, hominid groups would have encountered the large land mass of the Sunda Shelf, which then linked the tropical Southeast Asian archipelagos (including the island of Java) to the mainland. In our own century, dredging and fishing on areas of the Sunda Shelf that are now offshore have yielded Pleistocene fossils of large mammals. Although modern-day Longgupo appears inaccessible from these lowlands because of the eastward drainage of the upper Yangtze River, during the Pliocene the river drained southeast toward the Gulf of Tonkin. Like northern Pakistan, inland south-central China could have been reached directly from the south by ascending a broad alluvial valley. Indeed, paleontologist John van der Made of the Rijksuniversiteit Utrecht notes that in association with the hominids at Longgupo, subtropical species of horse and pig appear rather early at their northerly limits.

      If early Homo was able to disperse early and quickly across subtropical East Asia, why does it seem to arrive later in West Asia and yet later in Europe? For West Asia, the sites of 'Ubeidiya in the western Jordan Rift of Israel, and Dmanisi, southeast of Tblisi, Georgia, provide the most information on early occupation. 'Ubeidiya's 150 vertical meters of geological infill appear to have been deposited over a short period. The assemblage of animal fossils in the stratigraphic column suggests that much of the deposit dates to nearly 1.4 mya. The lower archaeological levels contain a range of stone flakes and tools, some having the standardized forms associated with contemporary assemblages in East Africa, but there are no biface tools. The upper levels, not greatly different in age, clearly have these Acheulian tools. As only a few isolated human teeth have been recovered from the site--only one in geological context-the hominids responsible for the stone tools remains unknown. Ofer Bar-Yosef of Harvard University interprets the sequence as the result of two separate hominid occupations within a relatively short time. Both groups shared an advanced core-tool technology; one group used biface tools, the other did not.

Figure 10. Early Homo managed natural resources to minimize dependence on specific territories. The simple strategy for "catchment" scavenging is evident in eastern Rift Valley archaeological sites older than 2 million years. A group of hominids using this strategy established a camp conveniently close to multiple primary resources, such as water, stone and animal carcasses (red marker in the foreground). The catchment becomes the area around the camp within which resources are distributed sparsely but still readily available (blue markers ). Archaeological remains at catchment camps are rich and diverse, whereas little material is found at outlying points. After 2 million years ago, Homo ergaster 's scavenging strategy became decentralized. This hominid ranged among more points within a larger "territory," a behavior that required more intimate knowledge of scattered resource points. Territory-site assemblages are thus smaller and more specialized, each reflecting localized resources. Rather than getting to know a landscape intimately, early Homo traveled widely among catchments (red markers in the background ), thus increasing its ability and chance to disperse widely.

      The deposits at Dmanisi were first excavated in the 1980s for their faunal remains. Later, an assemblage of simple flake and core tools was found and, finally, in 1991, a complete jaw of Homo erectus. The age of the hominid and the artifacts is debatable. The enveloping sediments have a paleomagnetic interpretation of about 1.8 mya and a potassium-argon determination on the underlying lava flow corroborates with 1.8 mya. Nevertheless, the hominid fossils and tools appear to lie in pockets having a younger paleomagnetic reading. A conservative biostratigraphic estimate for the remains is 1.4 mya, nearly identical with the dating of 'Ubeidiya. Combining the information from these two sites suggests that the Levant provided a corridor northward from Egypt and Saudi Arabia into Europe and northwestern Asia. Possibly, the eastern Mediterranean region has remained too rugged to support a Pliocene population of Homo, no matter how open the habitat. One other very enigmatic site plays into the timing of hominid dispersals to the Middle East. In the western Jordan Rift, the site of Erq-el-Ahmar has yielded a few probable core tools in a geological formation that antedates that of 'Ubeidiya. A paleomagnetic analysis indicates an age of about 1.8 mya, but some doubt remains about the date and the artifacts themselves.
      As for Europe, it is fairly clear that mountain ranges, large watercourses and generally rigorous climates inhibited any significant early dispersal. It is certain, however, that a precursor to Homo sapiens colonized much of central and western Europe with the aid of a mature Acheulian technology. The earliest of those fossils and assemblages date to 600,000 years ago. Paleoanthropologists have argued for years about the name of this hominid species. Recently, Philip Rightmire of Binghamton University has proposed that a number of similar cranial and dental fossils from across Africa and Europe can be assigned to the species Homo heidelbergensis, which was first applied to a jaw found at Mauer, near Heidelberg, Germany, in 1908. Homo heidelbergensis is best described as a large-brained form of Homo, but still with a massive face. This hominid seems to have emerged in southern Africa around an advanced Acheulian technology at sites such as Bodo and Elandsfontein. The combination of a large brain and an advanced technology may account for the hominid's quick dispersion into Europe, as both Homo heidelbergensis and the Acheulian appear at sites such as Mauer and Boxgrove (in far northern Europe) by 500,000 years ago. In Europe, Homo heidelbergensis is the hominid that begins to show traits similar to the Neanderthals about 250,000 years ago, starting a trend of anatomical specialization that probably leads to extinction less than 30,000 years ago. Neither Homo heidelbergensis nor the Acheulian follow the path of early dispersal across southern Asia.
      In Europe, as in Asia, recent finds and more penetrating age analyses bring up the possibility of African arrivals earlier than commonly hypothesized. Thus at Isernia la Pineta, in southern Italy, an elementary core-flake assemblage has a date of about 800,000 years ago, based on paleomagnetic and biostratigraphic methods. Even more intriguing are two Spanish sites where recent excavations have produced hominids and technological assemblages at an equally early age. At Atapuerca, in north-central Spain, core and flake tools accompany an early representative of Homo at greater than 780,000 years ago. Artifacts in a yet lower level of this locality may be significantly older. At Orce, in south-central Spain, excavators have recovered simple core and flake tools as well as purported hominid skull fragments and limb elements. Paleomagnetic analysis suggests an age of about 1.0 mya. Although hominid taxonomy and age determinations are tenuous at both localities, the Spanish sites represent possible evidence for an early species of Homo arriving from Africa before Homo heidelbergensis and its Acheulian technology. It is also possible that Homo reached Europe via Gibraltar at least once.

Early Dispersal and Homo sapiens

      If early Homo was able to leave Africa before 2.0 mya (and later), did these intercontinental quests contribute to the prehistoric distributions of Homo sapiens, our own species? The new evidence must be framed in terms of the ongoing debate over "multiregional" or "out-of-Africa" origins for Homo sapiens. The multiregional argument would use the new evidence to suggest that the early dispersals of early Homo set a complex stage for Homo sapiens to emerge at connected points across much of the Old World. For us, however, the new evidence suggests that both the early and later species of Homo had the ability to disperse across the continents. Given that Homo sapiens fossils appear to be much older in Africa than on any other continent, the answer seems obvious. Having emerged in Africa, Homo sapiens dispersed to Eurasia, replacing older populations of Homo.
      Although there is little question that Homo sapiens emerged in Africa, the date of emergence, the technological associations and the dates for its Eurasian dispersals are debatable. It appears that our species originated between 200,000 and 100,000 years ago, somewhere in sub-Saharan Africa. Some recent discoveries in Zaire of ancient and finely crafted tool types (such as barbed-bone harpoons) indicate that the technology associated with this emergence may have been very advanced indeed, resembling the much later Upper Paleolithic of Europe. In the Levant, where Homo sapiens is evident about 90,000 years ago, a more archaic Middle Paleolithic technology still held sway. Consequently, we may not yet say whether the European dispersal of Homo sapiens was associated with either its emergence or a new technology.

Figure 11. Dispersal corridors opened out of Africa and across the Middle East into South and East Asia during the late Pliocene. Corridors formed primarily along coastal land masses, the product of expanding polar ice caps and resultant sea-level drawdown. Incipient development of the Red Sea rift also established departure routes through the Middle East. The drop in sea level in island Southeast Asia would have connected Sumatra, Java and Borneo with the mainland. Evolutionary analysis of fossil species, including antelopes, pigs and cats, indicates that large mammals dispersed along these routes sporadically during the late Pliocene and early Pleistocene. New fossil finds suggest that early Homo arrived in Asia some 1.8 to 1.9 million years ago, after departing Africa at least 100,000 years earlier.

Conclusion

      In linking the early dispersal of early Homo with its emergence, we are describing a hominid very different from the australopiths, whose bipedal but still ape-like anatomy must have limited them to wooded locales. Thus the significance of an early dispersal to Asia is manifold. First, the climatic conditions of cool aridity that played a great role in the emergence of Homo itself also drew hominid populations out of Africa and into Asia. Emergence and dispersal are, to a great extent, a product of environmental change. Nevertheless, early Homo emerged with a radical, yet still generalized, set of characteristics that granted it ecological hegemony across the subtropical Old World. An early intercontinental distribution signifies a hominid not adapted to specific territorial conditions, but adapted to manage many local conditions through physical presence, technology and flexible social organization. Ironically, as the first species to use technology, early Homo colonized much of the subtropical Old World without the benefit of language, symbolic culture or individual consciousness as we know it.

Bibliography