Chapter 5
Savannas and human evolution
Savannas have helped to shape the evolutionary pageant of human history and the dispersal of our ancestors across the continents. Although there is considerable controversy over the causes and mechanisms of this evolutionary story, ecological, environmental, and genetic evidence suggests that human-like primates first arose in savanna sites within Africa. These habitats offered more open conditions of grassland and woodland that are believed to have stimulated the development of specific physical attributes and social behaviour leading to advanced and mobile societies. The early history of scavenger-hunter-gatherers and their migrations out of Africa into Asia, Australasia, and eventually into the Americas makes a compelling saga. Modern humans contended with climatic oscillations and changing, constantly challenging environments in their migratory spread over the Earth’s surface. The wide panoramic landscapes of savannas offered resources and opportunities that facilitated these extraordinary movements.
The origins of savannas
Savanna vegetation is characterized by a ground layer of grasses and associated herbaceous plants that forms a defining component of the landscape. Although the emergence of flowering plants (angiosperms) can be dated back to the early Cretaceous around 140 Mya, the grasses then formed only a small proportion of the total flora. The savanna landscapes as we define them today can be traced back to the evolution of grasses in the late Tertiary some sixty to seventy Mya. The large grass family (Poaceae, with around 11,000 species) became divided into those that flourished in more temperate climates (C3) and those that evolved to deal with warm climates (C4). It is likely that the grasses originated in the old Gondwana land masses as C3 shade-adapted plants but rapidly expanded from the Eocene (fifty-five Mya), spreading into open habitats by forty Mya. This was probably not a simultaneous event across the globe, but was first evident in South America (Eocene) and lastly in the Australian savannas (late Miocene-Pliocene).
Grasses with C4 metabolic pathways dominate tropical savannas today and account for possibly 30 per cent of global C-fixation. The physiological mechanism, referred to as a CO2-concentrating pump, enables the plants to have a high photosynthetic efficiency in warm climates where there are low natural CO2 levels. The C4 pathway proved to be a major evolutionary success, giving rise to grass plants that have been crucial to human development. These span a huge range of economically and culturally important plants such as maize, Zea mays first cultivated in Mesoamerica around 2,500 years ago, sugar cane (several species of the genus Saccharum), sorghum (from the sub-family Panicoideae), or the largest and fastest growing member of the grass family, bamboo (sub-family Bamusoideae) with around ten genera and nearly 1,500 species through the world. The same mechanisms have been identified in other, mostly herbaceous, plants such as sedges (Cyperaceae) and numerous other well-known plant families such as the daisy (Asteraceae). It has been widely accepted that the C4 pathways evolved as a response to declining levels of atmospheric CO2, making a CO2-concentrating mechanism a valuable adaptation to increase photosynthesis. However, this simplified explanation has been criticized and it appears that a combination of factors aided the rise of the C4 grasses. These included the influence of seasonal climates with strong dry episodes that led to water shortages and facilitated fires, especially in monsoonal regimes marked by a high incidence of lightning strikes generating fires at the stormy onset of the wet season. The influence of the increasing number of herbivores could also have played a part, since browsers would have reduced the leaf canopy of woody vegetation and limited re-growth. It is not clear whether the process of change over from C3 to C4 was a single evolutionary event or whether there were multiple mutations and possibly reversals as climate varied.
The rise of C4 plants was rapid in terms of geological timescales, starting in the late Miocene (five to eight Mya) and extending into the Pliocene (two to eight Mya). The same processes have continued since that time through the constant variations in climate. Direct plant fossil remains have long since decomposed and evidence for these changes is largely derived from C-isotopic data in organic residues and fossil soils. The transformation was accomplished rapidly across all the continents and the proportion of C4 grasses is believed to have risen from around 0 to 80 per cent in two to four million years. Since the takeover occurred globally at around the same time, it seems likely that there was a global reason, such as climate and levels of atmospheric CO2 concentration. The early grasses adapted to seasonality and warmth by developing drought-resisting features. The main growth points (meristems) are located at the ground surface or even within the soil, with increased root and decreased shoot growth, sunken stomata, and thick cuticles which help to resist fire. There is also strong evidence to indicate that the rise of grasses coincided with the spread of mammals since the grazers and browsers became adapted to a diet high in cellulose and silica-rich leaves and in this sense savannas represent an example of co-evolution. Other dryland plants also developed in response to seasonality such as drought-resisting shrubs. The typical savanna vegetation with high proportions of grasses and shrubs became evident by the Miocene around sixteen to eleven Mya.
Origins and movements of people
While there has been much debate about the part played by savannas in influencing the course of human evolution, the major outlines have broad agreement. The greatest numbers of the earliest hominins (belonging to the family Hominidae, with a single genus Homo) are to be found in sub-Saharan Africa, and it is generally agreed that the closest relations of modern humans evolved over large parts of that continent some two to five Mya. Evidence of these early ancestors has been found in eastern and southern Africa and also in areas as far apart as Malawi and Chad. The best known fossil evidence of the earliest remains has been discovered in Kenya, Tanzania, Ethiopia, and more recently in South Africa. The largest single fossil hominin assemblage found so far has been termed Homo naledi in the Rising Star cave system 50 km south of Johannesburg. It is thought that the open landscapes, but especially the more wooded savannas, offered opportunities for early pioneers to become hunter-gatherers and scavengers, and to range over extensive tracts of land. This is likely to have been the result of a combination of propitious environment (food resources and water) together with the development of advanced forms of social behaviour (integrated activities and the use of tools). It is believed that the first members of the genus Homo appeared 2.5 Mya. At around two Mya there appear to have been six human-like species, although the evidence is scattered, fragmentary, and open to constant re-interpretation. It was clearly a prolonged and complex process of evolution and while the mosaic of savanna environments provided a context, it is likely that the concentrations of early hominins were connected with the most favourable habitats having a proximity to water such as forest edges, wooded patches, and gallery forests.
From fossil evidence and information derived from a wide variety of methods such as isotopic dating, palaeomagnetism, amino acid analysis of organic materials, and genetic tracking, it would seem that the earliest primates and hominins ranged over a mosaic of environments. One authority commented that there is no evidence, in the early stages of human evolution, for an environmental change that would determine a spread into savanna. Nevertheless, some sites, such as the Olduvai Gorge in northern Tanzania, have been shown to reveal recurrent ecosystem variations. Open C4 grasslands changed relatively abruptly (over several hundreds to thousands of years) to closed C3 forests with periodic reversals. It does appear that there were climatic variations with markedly drier and wetter episodes in at least some of the areas of early hominoid activity, and this may have helped to trigger development.
It has been suggested that biological and human evolution, which proceeded in fits and starts, was at least partly occasioned by abrupt climatic changes stimulating or forcing new adaptive strategies. The open landscapes would have provided a broad spectrum of food resources whose exploitation required and has been correlated with increases in brain capacity. There is evidence of an increasing adjustment to the ground level through the development of bipedalism. This freed the use of hands and provided greater adaptability to the demands of seasonal rhythms. Over recent years there has been an emphasis in the literature on the growth and increasing sophistication of social systems, including the increasing power of communication through language. Savannas provided a range of environments that are likely to have stimulated such development. The evolution of language would have required an ability to communicate beyond a level possible with very small groups, and would in turn have accentuated social customs. Although some of more advanced human-like individuals such as Australopithecus afarensis seem to have been at home on the ground and in trees (for instance, the fossils found in the Hader area of Ethiopia), there are clear signs of later, greater adaptations to more open and savanna-like conditions. The detailed description of early hominins within the Olduvai Gorge, where there is a two-million-year sequence of lake and river sediments, illustrates this evolutionary panoply. H. habilis appeared in the Olduvai area around 1.9 to 2 Mya, with H. erectus—from 1.2 Mya—and H. sapiens very much later (see Figure 28). Broad-scale climatic changes continuously varied the landscape over these timescales. Some regions became wetter, such as the southern Sahara, as evidenced in the ancient stream channels now buried by sand. Equally there were periods that were markedly drier with extreme droughts when there would have been inadequate food resources.
28. The human ancestral tree and evolution from fossil hominins. Recent discoveries at Rainbow Cave 50 km south of Johannesburg have been named as Homo naledi and may represent some of the earliest of human remains.
The explosion of information in evolutionary genetics over the past few decades has offered further evidence on these origins and migrations. The genetic record provides a life story of living organisms, where the genomes of living individuals are passed down from ancestors and where ancient DNA is preserved in organic remains. It is claimed that it is now possible to catalogue the genetic differences between humans and our earliest relatives comprehensively, and to trace one layer of history on top of another. This evidence has confirmed that the first hominins evolved in Africa from six to seven Mya and that bipedalism had developed by between four and six Mya. The first humans appeared around 1.9 Mya and the speed of migration was then extraordinary since H. erectus has been recorded in East Asia by 1.8 Mya. The use of fire has been dated back around a million years in Africa. Anatomically modern humans appear over the last 200,000 years and differ from ancestors in their morphological structure and brain, in their social organization, and probably in language as well as in selected genetic changes. There is also evidence of a high level of genetic differentiation over time between relatively close populations. This has been interpreted as being the result of obstacles to dispersal such as mountain ranges or powerful water barriers. On the other hand the open savannas and coastal plains offered greater visibility, food and water resources, and accessible routeways. Integrating geographic information such as this with the genetic information will provide a powerful approach to interpreting genetic diversity in the future.
Dispersal and migrations
The dispersal of early humans away from their savanna-like origins in Africa makes up an epic narrative and took place over a remarkably short period of history—around 100,000 years. Current thinking indicates that some African fauna moved out in the warm climates of the last Interglacial around 110-130 Kya and that early hominins expanded about the same time (ninety to 130 Kya). A likely route out of the African continent would have been the region around the Awash Valley in Ethiopia, which is adjacent to the southern end of the Red Sea. This passage would have afforded a short crossing to the Arabian Peninsula and then along the coast to the Indian sub-continent. An alternative route further north might have taken early migrants into the Middle East and then east via the Caucasus, Black, and Aral Seas, and on into eastern China. The sea crossings would have been easier to pass at times when glaciations in the northern hemisphere had resulted in lowered sea levels. At different climatic phases, the migration routes would have been facilitated when increased rainfall promoted savanna-like vegetation in what are today semi-arid regions. Equally there is evidence of periodic dry episodes that would have halted expansion, cutting off routes with desert-like conditions.
There are differences of opinion as to whether there was one major route or multiple routes, and whether there was a continuous movement or pulses of migration in favourable periods. Whatever the route and whatever the number of movements, the savanna landscapes provided a range of favourable habitats for expansion. The migrations appear to have reached South-east Asia relatively rapidly and this region is claimed to be only second to Africa in the length of recorded settlement. Although it is not proven, it is suggested that the first hominins to reach South-east Asia were ancestors of H. floresiensis with similar physical proportions to those of H. habilis in the Olduvai Gorge. H. erectus appears in the record in Java and a few other sites only 100,000 years after their record in Africa. Anatomically modern humans are identified in the region from sixty to 50,000 years ago and current genetic evidence for H. sapiens tends to favour single migration out of Africa some fifty to seventy Kya.
The migratory movements passed through phases of expansion and bottlenecks that were largely driven or limited by climatic change, and shaped by the configuration of the land masses. By the time early migrants reached south and south-eastern Asia, many of the landscapes were drier than at present notably during the maxima of Pleistocene glaciations. Present-day Sri Lanka was attached to the Indian mainland and much of South-east Asia was joined into one land mass (Sundaland), separated from Australia and New Guinea (Sahul) by deep channels. These breaks in the landmass with deep, fast flowing currents, known as Wallace’s or Weber’s lines after the pioneer naturalists, severely restricted human and animal movements. Sea levels were estimated to be over 100 m lower than the current distribution, exposing land that is today under the sea. The extent of open savanna-like conditions was almost certainly much greater than the present day although it is probable that the main migration routes to the south-east were along coastlines and may have involved some of the earliest known seacraft (see Figure 29). The theory is that the expanses of open vegetation in Africa and elsewhere supported a proliferation of food resources including large herbivores and many smaller animals. This is believed to have stimulated the development of tools to catch and prepare food. Changes in the vegetation as shown from pollen data, indicate a transition from pine-alder associations at glacial maxima to broad grassland corridors with gallery forests along water courses as the climate became drier and more seasonal. This in turn attracted characteristic savanna animals such as elephants, rhinoceros, and deer.
29. Habitats crossed by early migrants in south and east Asia at low sea levels during the Late Glacial period, illustrating the extensive stretches of exposed land.
The area known as Sundaland in South-east Asia, exposed by the falling sea levels, was nearly twice the area of the present day land. Many of the islands making up present day Indonesia were joined together facilitating movement of early humans. Oscillations of sea level, the most recent of which was a rise around fourteen Kya, eventually cut off these pioneer groups which are likely to have been very small and they subsequently evolved in relative isolation. Many of the signs of these first venturers would be located below present-day sea level. Over higher land it appears that sites have been buried by the active volcanic debris typical of the island arcs of South-east Asia (for instance the catastrophic eruption of Toba volcano in Sumatra at around seventy-four Kya). The coastal route from Africa to South-east Asia and then on via the island arcs to Australia is probably the principal migration passageway. However, it is likely that there were multiple expansionary movements of peoples taking advantage of the open landscapes of woodland-scrub-grassland that existed over large parts of Asia in the moister climatic phases.
Passage to Australia
Although the crossing from the nearest South-east Asian islands appears straightforward with relatively short distances between them, the straits are frequently deep and treacherous. There is little doubt that early migrants did cross significant stretches of sea but they were believed to be mostly small groups and often became isolated. Despite this, some genetic evidence proposes that the first group to reach present day Australia was a substantial number (possibly around 1,000), which would represent a surprisingly large migration. Movements of this size would contradict the idea of very small, gradual, incremental, and episodic shifts of people. The closest part of Australia to places such as New Guinea would only have been a little over the line of sight at the times of the lowest sea levels in the past. It has been suggested that there might have been indications of land, such as the appearance of land-based birds and a dust haze suggesting a landmass. There is evidence to show movement between the islands (and later out into the Pacific), so the early colonizers might have acquired the ability to cross by sea. There is no suggestion that H. erectus ever reached as far as Australia, and there are several different dates put forward for the first landfalls although they are likely to have been between forty-five and seventy Kya. The first settlers in present day Australia would seem to have been in Arnhem Land, where artefacts have been found about 50 km inland dating to sixty-one Kya ± 10. However, the initial landing sites are almost certainly under the sea today and were maybe 200 km further north of the present coastline. There were numerous food resources including a number of very large animals (such as the now extinct giant flightless birds and kangaroos). These are vividly depicted in the cave art for which the indigenous areas are famous.
The early migrants have been described as supernomads. They were able to cover extremely large areas remarkably rapidly. They appear to have adapted to the varied landscapes, particularly the more open savannas that resulted from the oscillations of climate in the post-glacial period. For instance there are palaeo-records showing widespread flooding from eighteen Kya, with shrub vegetation seemingly predominant at around thirteen Kya and forest areas becoming dominant from ten Kya. The present day monsoonal regime of northern Australia appears to have developed by fourteen Kya and extensive savanna (woodland and grassland) by around eleven Kya. Most of the coasts, open landscapes, and forest were occupied to some degree by thirty to forty Kya. Adaptations to the semi-arid landscapes took longer but seasonal rainfall and more moist periods in the past appear to have allowed access and even settlement in favourable locations. Essentially these colonizers were highly mobile and adaptable to the changing climatic regimes. The savannas seem to have acted as a conduit for the migrants who utilized the mosaic of forest, grassland, and desert—as long as there was water, food, and wood for fuel. Savannas seem to have been burnt from the earliest times (e.g. by travelling groups carrying the traditional firesticks), and this may have exacerbated the increasing dryness. For example, Lake Eyre shows a record of a tree-shrub-grass mosaic being transformed into a desert landscape some sixty to forty-five Kya.
Passage to the Americas
Unlike the model described for the Old World, savannas played very little part in the settlement of peoples across the continent, although they may have played a part in promoting movement. The widely accepted theory is that the initial colonization of the Americas took place via the Bering Strait from Asia at times when the last glaciations had locked up much of the oceanic water and sea level drop meant that a wide causeway (Beringia) existed from what is now eastern Siberia to Alaska. Although there are opinions that some migration was earlier, and some have suggested that it was partly trans-oceanic, most authorities agree that the dispersal into what is now North America took place after around seventeen Kya. The movement of small groups of people out of Africa, through the Middle East, then on to the north-east of Asia, and eventually into Beringia was inevitably a slow and tenuous affair slowed by distance, climatic changes, and environmental barriers. The onward movements spread over the continent may have reached as far south as Mexico, but as far as the South American continent is concerned the main direction would have been a southerly route via the western coast or along the flanks of the Rockies.
Early migrants passed down through Central America, probably following the more accessible routes via higher land or along the coast. There is only very scattered evidence of these first groups, and the earliest dated remains (rock shelters, cooking hearths, maize cobs) are arguably from nine Kya. At lower altitudes where the climate was more typically tropical, there is a nearly continuous series of savanna patches on elevated plateaus, and less commonly over lowland plains closer to sea level. These appear to have been utilized and settled at later stages in the migrations. There is evidence of severe droughts and animal extinctions similar to the movements through Africa and Asia, and the impact of climatic changes appears to have caused analogous bottlenecks to migration. The pioneer colonizers of the tropical parts of Latin America appear not to have occupied the savannas as much as the forest environments or temperate uplands. The great civilizations that grew up in the continent appeared at a later date. They tended to be located on higher ground such as the Mexican plateau (Aztec and their predecessors), and followed the line of the Rockies-Andes cordillera (Inca and pre-Inca groups on the Pacific coastal plains). These early groups of peoples moved very rapidly over the higher land to the south of the continent. Other groups settled in the humid Atlantic coastal plains of Central America (Olmec and Maya) and, perhaps by following the coastline, in Amazonia. There are scattered savanna tracts throughout the rainforest with seasonally flooded formations along river courses (várzeas), on Marajó Island and coastal areas in the mouth of the Amazon. The densest occupation by indigenous groups appears to have been in the eastern forests, particularly alongside rivers. The early settlers over the savannas would have needed access to water resources. Recent research has shown distinctive genetic episodes for these initial movements of people and it is believed that some of the pioneer Amerindian groups became isolated quite early in the migrations.
Although the tropical savannas were probably not the principal routeways in South America, there is evidence that some of the colonizing groups found ingenious ways of utilizing more favourable sites where there was water. There is evidence of extensive occupation with drained and raised fields in western moist savannas such as the Llanos de Mojos similar to the land uses in a number of lowland Mayan sites. Around the Amazonian fringes, Amerindian groups like the Kayapó, appear have deliberated transferred soil over short distances from forest to savanna in order to plant different species and diversify the sites. The open landscape was probably used more for access and hunting. Nevertheless, the Central American, llanos, and cerrado landscapes were not settled by substantial groups until much later than the African model and remained relatively sparsely populated until colonial times and later.
There have been remarkable stories of migration in almost every historical era and it has been suggested that all the major language groups were in existence by around 30,000 years bp. The savanna biome has evolved relatively recently in geological history and the rise of humans has only occupied a very short fragment of this time. As climate has varied so the forest-savanna and savanna-desert boundaries have shifted to and fro with oscillations of rainfall and temperature. The arrival of the first humans accentuated the impact of fire, constantly gnawing at the forest frontier and renewing the grass-dominated plant life. Forms of agriculture appeared around fifteen to five Kya and developed independently in several regions following the end of the last glacial period in the northern hemisphere. The area from Egypt to western Asia through to India probably saw some of the earliest sedentary forms of agro-pastoralism, often based on primitive grasses. The Fertile Crescent in the Middle East is credited with the development of wheat, the mid-Nile with sorghum, while South-east Asia witnessed the cultivation of rice. Pastoralism also has an early history with domestication of cattle and possibly the gradual incorporation of wild animals into a more domestic way of life.