Abstract

With the onset of the modern ice age, climatic changes in Africa caused grasslands to expand and forests to contract. This environmental shift appears to account for the evolution of Homo about 2.4 million years ago (Ma). Evolution could not have established the mode of development that produces the large brain of Homo until human ancestors were fully terrestrial. This mode of development produces the long interval of physical helplessness that distinguishes humans from all other mammals; it yields infants that cannot cling to mothers. The more mature infants of apes can cling to their mothers, allowing the mothers to climb trees. Gracile australopithecines, from which the genus Homo evolved, retained an apelike pattern of development, arboreal adaptations, from their ancestors. Thus, even females with infants could have been adept tree climbers, and the need to avoid predators must have required that gracile australopithecines be habitual climbers. In contrast, the small pelvic dimensions and large brain of early Homo point to delayed development, helpless infants, and a totally terrestrial mode of life. The origin of Homo, about 2.4 Ma, appears to have resulted from the onset of the recent ice age, when climatic changes in Africa caused savannas to expand at the expense of woodlands. These changes must have had a severe impact on australopithecines, as they did on other groups of mammals. In particular, they should have compelled many australopithecine populations to abandon the arboreal activity that had maintained evolutionary stability for more than 1.5 million years (m.y.). The resulting crisis conditions presumably caused the extinction of many populations, but the evolution of a huge brain through delayed development was now possible for the first time. Selection pressure for superior intelligence fostered the development of advanced social behavior and tool manufacture, which offset the problems created by helpless infants and the loss of arboreal refugia and food resources. African antelopes experienced parallel changes, with forest-adapted species suffering heavy extinction and a variety of new species coming into existence.

Introduction

Some adaptive breakthroughs in the history of life have resulted from natural selection for improved adaptation without the influence of environmental change. Other major evolutionary transitions have resulted from selection pressures imposed by a changing habitat. Thus, not only do severe environmental changes alter the biosphere by causing migration and extinction, they also stimulate evolution. In fact, a species can, in a sense, experience both extinction and successful evolution as a result of environmental crisis. A species as it was constituted for 10⁶ to 10⁷ generations may die out and yet have one or more of its populations emerge as a distinctive new species adapted to the new conditions.

The onset of the recent ice age, at about 2.5 Ma, transformed habitats in many regions of the world. A variety of circumstantial evidence suggests that at this time a particular kind of environmental crisis in Africa caused the extinction of one or more species of the human family but also triggered the evolution of the modern human genus, Homo (Stanley, 1992). Homo evolved from a species of the genus Australopithecus. This ancestral genus, having been confined to Africa, could not have escaped the environmental changes that pervaded this continent.

The early evolution of Homo was of unique significance in the history of life. Most importantly, it entailed a marked increase in brain size. The early Homo fossil skull 1590, which belonged to a child about 6 or 7 years old, would have grown to have an adult cranial capacity well in excess of 800 cm³, about twice that of a male belonging to the gracile australopithecine species from which early Homo evolved.

The term ''gracile" means slender, and applies specifically to the jaws, teeth, and skull, which in both of the known gracile species Australopithecus afarensis and A. africanus were less heavily developed than in the "robust" species, which many authors now segregate into the genus Paranthropus. The robust forms were almost certainly evolutionary offshoots of the line of descent that led to Homo. Most experts now regard A. africanus as the gracile species most likely to have given rise to Homo, in part because the known range of this South African form extends to about 2.3 Ma, more than half a million years beyond that of A. afarensis, whose fossil record is confined to northern Africa. In any event, the two gracile species were rather similar in general morphology and presumably therefore in mode of life.

It is important to recognize that the gracile australopithecines existed for more than 1.5 m.y., from about 4 Ma to perhaps 2.3 Ma, without experiencing appreciable evolutionary change. Not only did their cranial capacity remain only slightly above the level of a chimpanzee, but their postcranial anatomy experienced approximate evolutionary stasis as well (McHenry, 1986). They were obviously successful, well-adapted creatures. It has long been understood that australopithecines walked bipedally. Their pelvic configuration is much more human than apelike in form, and their hindlimbs also in many ways resemble those of Homo. Furthermore, fossil bipedal footprints in Tanzania were formed at about 3.7 Ma, long before Homo or robust australopithecines evolved. By default, these tracks are attributed to gracile australopithecines (Leakey and Hay, 1979).

Although there remains no doubt that australopithecines moved bipedally on the ground, the past decade has seen the emergence of abundant evidence that tree climbing was also part of their normal behavioral repertoire (Stern and Susman, 1983; Susman et al., 1984; inter alia). The bones of their forelimbs display numerous traits that evolved as arboreal adaptations in their ancestors. The bones of their hindlimbs were more human in form, but nonetheless retained certain traits that would have enhanced the ability to climb while at the same time restricting performance in walking and running to a level below that of modern humans. The habits of extant primates indicate that australopithecines should have been required to use their appropriate traits in order to avoid predators. The australopithecines' division of activities between terrestrial and arboreal habitats accounts for the failure of their postcranial morphology to evolve appreciably in a human direction for some 1.5 m.y.

In a less direct way, semiarboreal behavior can also account for the gracile australopithecines' failure to evolve appreciably larger brains during the same interval. In fact, it was impossible for the large brain of Homo to evolve until obligate arboreal activity had been abandoned. This restriction related to the developmental mechanism by which the large brain evolved: extension into the postnatal interval of the high in utero rate of brain growth, which in all primates maintains brain weight at about 10% of body weight. In monkeys and apes, this rate gives way to a much lower rate soon after birth. In humans, however, cranial development is delayed, so that the high fetal rate of brain growth is projected through the first year of life after birth. This produces a cranial capacity for a 1-yr-old human infant that is more than double that of an adult chimpanzee. Figure 14.1 illustrates the differing developmental patterns for apes and humans.

Figure 14.1

Differences in brain growth between apes and humans, and similarities between apes and australopithecines. Solid symbols represent males; hollow symbols, females. Circles represent chimpanzees; triangles, orangutans. The Phase I slope is shared by all (more...)

The reason that human ancestors could not evolve the large brain of Homo until they totally abandoned arboreal activity is that the developmental delay that yielded the large brain was linked to other aspects of maturation. A general retardation of development produced the large brain by projecting the high fetal rate of brain growth into the postnatal interval. In contrast to infant chimpanzees and orangutans, which are mature enough to cling to their semiarboreal mothers soon after birth, infants of early Homo had to be carried. It follows that, because a mother cannot climb while carrying an infant, a totally terrestrial life was a prerequisite for the evolution of Homo.

As it turns out, we can infer from fossil evidence the body and brain sizes of newborn and adult australopithecines. From these data, we can show that gracile australopithecines were apelike in their pattern of development: their infants were not physically helpless and should therefore have been able to cling to climbing mothers. Figure 14.1 illustrates how the pattern for gracile australopithecines resembles that for modern apes. On the other hand, comparable data for early Homo indicate a modern human pattern: infants were helpless (Stanley, 1992).

There is extensive evidence that at about 2.5 Ma, African forests shrank at the expense of grasslands, which are adapted to drier conditions. It has been suggested that these major vegetational changes that swept across Africa caused evolutionary turnover within the human family (Vrba, 1975, 1988, inter alia). In fact, the changes were precisely the kind that could be expected to have forced australopithecines to abandon habitual arboreal activities (Stanley, 1992). The result would have been extermination of many populations at the hands of large predators, but also provision of the opportunity for evolution of a large brain because physically helpless infants were now tolerable. The environmental changes should also have engendered strong selection pressures for brain expansion, given the need of hominids to cope with predation without escaping into trees and to replace tree-borne food materials.

Development In Apes, Humans, And Australopithecines

Estimates of brain volumes, body weights, and pelvic dimensions lead to the conclusion that the infants of early Homo were physically helpless, like those of modern humans (Stanley, 1992). A key fact underlies calculations leading to this conclusion: all primates adhere to the same general curve when their brain weights are plotted against their body weights for the fetal interval: brain weight constitutes roughly 10% of body weight (Holt et al., 1975). Figure 14.1 shows how brain growth in humans departs from that of apes in the postnatal interval. In apes, the fetal (Phase I) slope gives way to a much lower (Phase II) slope soon after birth, so that there is only modest postnatal encephalization. In humans, the inflection is delayed to an age of about one year, so that an infant this age still has a brain that constitutes nearly one-tenth of its body weight. The remarkable fact is that the brain of a one-year-old human is more than twice as large as that of an adult chimpanzee or orangutan.

Several observations indicate that a general retardation of development produced the delay in the Phase I-Phase II transition that yielded the large brain of modern humans. Overall retardation is evident in the physical helplessness and slow maturation of human infants compared to the offspring of apes. The evolution of highly immature infants in human evolution represented a profound ecological sacrifice. Offspring that must be carried, fed, and protected for years occupy parents' time that could otherwise be spent in important activities such as acquisition of food. Such infants also complicate the avoidance of predators. For natural selection to have produced this deleterious developmental pattern, there had to be some overriding benefit. Encephalization was clearly the change that provided this benefit: not only did it have great adaptive value, but it was as profound a change as the delay in development. In degree of brain expansion immediately after birth, modern humans rank first among mammals (Count, 1947), just as they rank first in the length of their postnatal interval of physical helplessness (Krogman, 1972).

No such developmental delay characterized gracile australopithecines. Fossils reveal that these animals closely resembled apes rather than humans in pattern of brain growth. It is most meaningful here to consider the pattern for males, because their mean birth size is larger than that of females, so that its maximum value can be estimated from the size of the female pelvis. Pelvic inlet breadth is the dimension that limits cranial size and therefore body size for neonates. This dimension for the pelvis of the famous "Lucy" skeleton, which represents Australopithecus afarensis, indicates that male birth size for this species approximated that for a chimpanzee or orangutan (Tague and Lovejoy, 1986). Other fossils show that adult male body size averaged at least 45 kg, not far from the mean for chimpanzees (McHenry, 1991), and that adult male brain size (averaging about 480 cm³) was only slightly above the chimpanzee mean (Figure 14.1). Apes do not give birth to neonates as large as their pelvic dimensions would allow. Thus, they do not develop brains as large as they might, even without postnatal extension of the Phase I portion of the brain-body growth curve. The probable reason is that apes' forelimbs are so heavily occupied in locomotion—arboreal climbing and terrestrial knuckle walking—that extensive tool use is impossible. Encephalization beyond the present level is therefore unwarranted, given the costs that accompany brain expansion: the demographic sacrifice that a lengthened gestation time would entail, for example, and the high energy expenditure required for growth of brain tissue. As a result, there is no reason to believe that the slight postnatal extension of the Phase I slope in gracile australopithecines resulted in helpless infants that could not cling to climbing mothers.

One might ask why natural selection should not simply have expanded the australopithecine brain without being required to retard development in general, with all the attendant problems. The primary answer is that an organ such as the brain does not develop in isolation, but is morphogenetically linked to other anatomical systems. As a result, a general delay in development was by far the simplest mechanism for brain expansion. All that was needed was prolongation into the postnatal interval of a pattern of development that was already in place. The complexity of other potential evolutionary mechanisms was so great that the probability of their occurrence was very low. A likely second problem with other mechanisms would have been their inability to expand the brain dramatically soon after birth. Delayed development offered the key advantage of producing a large brain during the first year, thus permitting the immense human learning process to proceed rapidly at a very young age.

The Life of Gracile Australopithecines

Even if clinging neonates permitted australopithecines to engage in arboreal activity, habitual climbing would have been possible only if the adults possessed appropriate adaptations. In fact, numerous of their morphological traits indicate that although these animals were adapted for bipedal locomotion on the ground, they were nonetheless much more adept climbers than modern humans. Furthermore, as I will explain below, one can make a strong case that australopithecines were compelled to put their climbing abilities to use in their everyday life.

Arboreal Traits

Modern humans can climb trees better than many members of advanced civilizations recognize, not only by shinnying but also by what amounts to walking up tree trunks. Members of certain Malaysian tribes are excellent barefoot climbers, gripping a tree trunk with a hand on each side, taking small steps upward by applying their splayed feet to the trunk, and then regripping with the hands at a higher level (Wood-Jones, 1900; Skeat and Blagden, 1906). Sometimes they climb so rapidly in this manner as to be described as "running" up trees. A variety of attributes would have made australopithecines more adept at these activities than modern humans are (Figure 14.2).

Figure 14.2

A gracile australopithecine "walking" up a tree. Key adaptations rendering australopithecines superior to modern humans in this activity are long arms relative to leg length; powerful wrists and hands with long, curved toes; ankles capable of great (more...)

A trait of australopithecines that would have given them a substantial advantage over modern humans in "walking" up trees was their larger ratio of arm length to leg length. This is reflected in the humerofemoral index, which is the ratio of upper arm length to upper leg length (Figure 14.3). The australopithecines' relatively long arms permitted the upper torso to tilt backward, so that gravity contributed more to the strength of the hands' grip (Cartmill, 1974; Jungers and Stern, 1983). This principle is the one employed by a repairman who climbs a telephone pole by looping a strap around the pole and his waist.

Figure 14.3

Relative lengths of the humerus (upper forelimb bone: left-hand member of each pair) and femur (upper hindlimb bone) for a pygmy chimpanzee, human pygmy, and Lucy (female Australopithecus afarensis). The length ratio for Lucy is intermediate. (See (more...)

The australopithecine forelimb exhibits additional traits that would have enhanced climbing ability. The glenoid cavity, the socket of the shoulder blade that receives the head of the humerus, is more upward directed than in modern humans (Vrba, 1979; Stern and Susman, 1983; Stanley, 1992). This is advantageous for suspensory activity.

Various skeletal features indicate that the australopithecines' wrists and hands were more powerful relative to body size than those of modern humans. In addition, their finger bones were long and curved, resembling those of chimpanzees (Figure 14.4). In fact, their hands were more apelike than human in form (Stern and Susman, 1983; Aiello and Dean, 1990). All of these traits would have served australopithecines well in climbing.

Figure 14.4

Strong curvature of a proximal phalanx (finger bone) from the hand of an individual of Australopithecus afarensis (AL 333-W4). This bone resembles the equivalent bones from the fingers of Pan (chimpanzees).

The hindlimbs of australopithecines also bore toes that were comparatively long and curved (Stern and Susman, 1982; Latimer and Lovejoy, 1990). Although these animals could not have grasped tree trunks or branches with their hind feet in the manner of apes, which have highly prehensile big toes, they were nonetheless better equipped than modern humans to grip arboreal substrata with curled toes. In addition, their ankles allowed for much greater upward flexure of the foot than ours (Latimer et al., 1987), which would have been very useful for vertical climbing (Figure 14.2). Finally, the relatively small body sizes of gracile australopithecines would have facilitated climbing by offering a higher ratio of strength to weight than characterizes modern humans. New estimates suggest that females averaged about 30 kg and males about 45 kg (McHenry, 1991).

In summary, australopithecines were certainly better climbers than modern humans. There is no question, however, that they were also less proficient at arboreal activity than modern apes. Their morphology suggests ability for upright climbing, but not extensive acrobatic activity of the sort engaged in by chimpanzees and orangutans. Once in a tree, they probably walked bipedally on limbs, gripping branches with their forelimbs for support.

The Arboreal Imperative

It is clear that the traits described in the preceding section evolved as arboreal adaptations in australopithecine ancestors, although unfortunately the fossil record of African primates for the Late Miocene and very early Pliocene is too poor to reveal the identity of these more heavily arboreal predecessors. The upright climbing posture of australopithecine ancestors was retained during the evolutionary transition to activity on the ground; there is no evidence that knuckle walking of the sort employed by modern apes had any place in human ancestry. We would not predict that a transition from a strictly arboreal mode of life to an essentially terrestrial one would occur instantaneously on a geological scale of time. Whether the change took place gradually or in steps, there should have been intermediate stages represented by taxa that divided their time between activities in trees and activities on the ground. The morphology of the australopithecines suggests that they were such taxa.

As Figure 14.1 shows, the australopithecines' pattern of development was apelike, so that the ability of infants to cling should have permitted a mother to climb. One might nonetheless hypothesize that the australopithecines had converted to a totally terrestrial life but, through some kind of evolutionary inertia, retained inherited arboreal traits. There are major difficulties with this idea, however. Traits that had evolved as arboreal adaptations persisted for a total of about 3 m.y. in the australopithecine complex of species. The robust australopithecines inherited from the graciles most of the "arboreal" traits that the latter themselves had inherited. In general, the australopithecines remained intermediate between apes and humans in a variety of locomotory features. Their forelimbs possessed many apelike traits. Their hindlimbs, having been occupied extensively with bipedal locomotion on the ground, were more human in form but seemingly remained compromised by residual arboreal activities: It is highly unlikely that evolution would have failed to improve terrestrial locomotion by eliminating these deleterious traits had they not been employed in essential arboreal activities.

Short legs were one of the traits that natural selection for improved locomotion on the ground might be expected to have eliminated if stabilizing selection were not maintaining them because of their value in climbing. The short legs of australopithecines relative to body weight must have reduced endurance in bipedal locomotion by increasing the number of strides per unit of distance traversed (Jungers and Stern, 1983). Similarly, the relatively long toes of australopithecines, though useful for climbing, would have reduced speed and endurance in running. Toes of some minimum length, apparently approximated in modern humans, are necessary for gripping the substratum and providing balance. Longer toes, however, lengthen the moment arm about which the body's center of gravity must rotate upward and forward during the so-called toeoff stage of running (Stanley, 1992; Figure 14.5).

Figure 14.5

Illustration of the reason that lengthening of human toes would impair running. Arrow shows the force that leg muscles apply to the long tendon of the big toe in elevating the body during the toe-off stage of running. The tendon attaches to the distal (more...)

It is not difficult to understand why australopithecines should have retained compulsory arboreal activities. Their dental morphology has been taken to indicate that fruits and seed pods formed a large part of their diet (Kay, 1985), and climbing would have expanded the range of available food items of this type. Even more important, however, should have been the need to elude large African predators.

The fossil record reveals that the group of large mammalian predators that inhabited the African continent during Early Pliocene time was much like the one that exists today. In fact, the Early Pliocene fauna was slightly more diverse. In addition to the single living species of lion, leopard, and cheetah, and the three living species of hyenas, there were at least five species that are now extinct: two additional hyenas and three sabertooth cats. It is inconceivable that australopithecines could have withstood the predation pressure of this formidable array of carnivores had they not habitually climbed trees. Significantly, large terrestrial herbivores in modern Africa closely resemble their predators in running speed. The difficulty that modern carnivores encounter in capturing these speedy prey is indicated by two facts: first, they focus heavily on young, old, and sick animals; and second, the availability of food generally limits their population sizes (Kruuk, 1972; Schaller, 1972). Australopithecines, having been even slower than modern humans, would have been no match for predators. Lacking fire and stone weapons, they would also have had little ability to ward off attacks on the ground.

Australopithecines in trees would have faced few effective predators—perhaps only leopards and the false sabertooth, Dinofelis. (The true sabertooth cats, having had long, fragile canine teeth, are thought to have specialized on pachyderms; summary by Marean, 1989). Furthermore, leopards are solitary, territorial predators rather than group hunters, so that their density and hunting prowess are both relatively low. A treed australopithecine with a sharp stick might have warded off a leopard whose forelimbs were occupied in clinging to a branch, but australopithecines confined to the ground would frequently have found themselves within the ranges of several species of social predators with excellent night vision and a preference for nocturnal hunting.

Modern primates that resemble australopithecines in body size offer a test of the idea that australopithecines would have needed to climb trees in order to reduce predation pressure. Chimpanzees and baboons both habitually employ trees as arboreal refuges in two ways. They sleep in trees at night, and they flee into them during waking hours when threatened by predators. Male baboons, with their formidable canine teeth, have been seen to face down leopards, yet when lions are in the vicinity, trees are as important a limiting resource for baboons as are food and water (Devore and Washburn, 1963).

In addition, accumulations of gracile australopithecine bones in South African cave deposits appear to be the products of predation (Vrba, 1980; Brain, 1981). These primates, together with baboons, greatly outnumber ungulates and are represented primarily by cranial remains. The implication is that the australopithecines were under heavy predation pressure. In summary, for more than 1.5 m.y., australopithecines retained traits that made them much better climbers than modern humans. The fact that evolution failed to rid them of these traits despite the fact that some of the traits were deleterious to terrestrial locomotion suggests that stabilizing selection was maintaining the traits. The source of this stabilizing selection is readily found in the need to climb trees frequently to feed and especially to avoid numerous species of fast, powerful, group-hunting predators.

The Nature Of Early Homo

The taxonomy of early representatives of the genus Homo is controversial, in part because of a patchy fossil record. "Early Homo" is a convenient label for fossil representatives of the genus older than Homo erectus, which ranges back to about 1.6 Ma. Traditionally, early Homo specimens have been assigned to the single species Homo habilis, which many workers now judge should be divided into two or more species. Details aside, it is clear that by about 2 Ma there existed some members of the genus whose brain capacities were at least twice the average for a gracile australopithecine. In addition, some members of early Homo had a pelvis that was not compressed from front to back, like that of a gracile australopithecine, but was instead remarkably like that of a modern human, except in having a smaller inlet, which required that babies be born at a smaller size than ours. Two known femora (thigh bones) of early Homo that are dated at about 1.9 Ma are also well within the length range for modern human females (Kennedy, 1983). In contrast, Lucy's femur is considerably shorter even than that of a female pygmy (Figure 14.3). All of the early Homo fossils mentioned above are quite similar to the equivalent skeletal parts of Homo erectus and modern humans, indicating a high level of adaptation to terrestrial locomotion.

The pelvic dimensions of early Homo indicate a small birth size; yet the remarkably large early Homo skull KNM-ER 1590, representing a small child, would have expanded into the Homo erectus range in adulthood (perhaps exceeding 900 cm³ in cranial capacity). The enormous amount of brain growth between birth and adulthood in early Homo would have required a considerable extension of the Phase I interval of growth into the postnatal interval (Stanley, 1992). Thus, early Homo could not have been an obligate tree climber: its infants could not have clung to their mothers.

Climatic Forcing

The oldest known fossils representing big-brained Homo are dated at 2.4 Ma (Hill et al., 1992). The oldest known manufactured stone tools also date to about this time, and these are customarily attributed to early Homo (Harris, 1983). The tools are simple flakes that represent the so-called Oldowan culture. The youngest gracile australopithecines are not precisely dated, but the famous Taung skull of Australopithecus africanus is now dated at about 2.3 Ma (Delson, 1988).

Gracile australopithecines had existed for at least 1.5 m.y. without experiencing appreciable evolutionary change by the time that one of their populations turned into Homo. I have argued above that (1) their postcranial morphology was straitjacketed in an adaptive compromise between terrestrial and arboreal activities, and (2) obligate arboreal activity also prevented them from becoming encephalized appreciably above the level of an ape. It is reasonable to conclude that some kind of environmental change would have been required to end their nearly static evolutionary condition. In particular, what should have been required was a change that caused at least one population to abandon habitual arboreal activity.

As it turns out, the onset of the recent ice age at about 2.5 Ma produced exactly the kind of environmental change in Africa that could be expected to have shifted australopithecine behavior in the appropriate direction. Africa became markedly drier, like many other regions of the world at this time (see review by Stanley and Ruddiman, Chapter 7, this volume). As a consequence, forests shrank and grasslands expanded. Fossil pollen reveals that in the Omo Valley region of Ethiopia, climates were warmer and moister than today before about 2.6 to 2.4 Ma, but cooler and drier than today thereafter (Bonnefille, 1983). Similar changes are recorded from Algeria, Chad, and Kenya (Coque, 1962; Conrad, 1968; Bonnefille, 1976; Servant and Servant-Vildary, 1980). Carbon isotopes in soils provide more detailed evidence of this change (Figure 14.6). Samples from a large number of hominid sites reveal no canopied forests after about 2.5 Ma and also document the first occurrence of wooded grasslands at about this time (Cerling, 1992).

Figure 14.6

Stable carbon isotope composition and inferences about floral composition for paleosol carbonates from East African fossil localities. Isotopic compositions for modern biomes are shown above. Figure after Cerling (1992).

That the floral changes in Africa had a profound effect on mammals is well established. Close to 2.5 Ma, numerous species of antelopes that had adapted to forest conditions suffered extinction, and during the next few hundred thousand years, there appeared a variety of new savanna-dwelling species, most of which survive as elements of the modern African fauna (Vrba, 1974, 1975, 1985a; Vrba et al., 1989). Micromammals underwent similar changes (Wesselman, 1985).

It has been suggested that the climatic changes may also in some way have promoted evolutionary turnover within the human family (Vrba, 1975, 1985b; Vrba et al., 1989). The changes of behavior and ontogenetic development that I have attributed to the origin of Homo suggest a particular mode of climatic forcing. Before the shrinkage of forests, troops of australopithecines probably occupied woodlands, which consist of groves or copses of trees separated by small areas of grassland. They could not have climbed well enough to have moved into and through the tall canopies of dense forests. Presumably, they used groves as home bases, sleeping in trees and occasionally feeding in them during the day (Rodman and McHenry, 1980). They may well have spent most of their waking hours on the ground, but only by remaining close enough to the home base to seek arboreal refuge when predators threatened. Modern baboons use trees in this way, even though they feed primarily on grass.

Saddled with the low intrinsic rate of natural increase that characterizes species of large primates because of solitary births (as opposed to litters) and lengthy generation times, australopithecine populations could not have sustained themselves in the face of heavy predation without arboreal refugia. Their relatively slow speed and weak natural defenses, in combination with their lack of both controlled fire and manufactured stone weapons, would have created intolerable predation pressure. They would have been easy targets for the multispecies guild of large, group-hunting terrestrial predators that, in the Pliocene as today, would have sought out as preferred prey animals that were easiest to catch. Even a grove of trees that initially served well as a refuge could not have sufficed indefinitely. In time, a troop would have exhausted food resources within close range of any home base. It would then have been required to move across grassland at the risk of suffering predation.

Before 2.5 Ma, woodlands were widespread and numerous groves of trees were separated by narrow zones of grassland. When forests shrank and fragmented with the onset of the ice age, however, many populations of australopithecines must have suffered a devastating intensification of predation pressure. Shrinking groves of trees offered smaller stores of food, which necessitated more frequent migration, and expanding grasslands increased the risk of predation by lengthening dangerous journeys. Presumably, many populations suffered extinction. Others may have survived for a time in areas that continued to support woodlands of moderate extent.

Widespread replacement of woodland habitats by grasslands is also exactly the kind of environmental forcing factor that could be expected to have obliged some populations to abandon habitual arboreal activity. Such a restriction of behavior automatically opened the way for encephalization through evolutionary extension of Phase I growth into the postnatal interval: physically helpless infants, though ecologically problematical, were now tolerable because mothers no longer climbed trees. Overriding the problems of raising highly dependent offspring, coping with predators, and losing arboreal food resources were the profound advantages of brain expansion—especially the ability to offset relatively weak physical attributes with innate cunning, advanced cooperative behavior, and sophisticated weaponry. These advantages of encephalization applied not only to avoidance of predators but also to development of hunting prowess that expanded trophic resources on the ground.

An important aspect of this scenario is that the first step was a simple change in behavior—one that amounted to a reduction of the preexisting behavioral repertoire. The result was that powerful natural selection pressures were brought to bear, so that major morphological changes ensued. Furthermore, the evolutionary retardation of development that produced encephalization was a relatively simply change, in that it represented only a modification of timing, not the origin of an entirely new pattern of development. This is not to say that the brain changed only by expanding. There was also a reorganization of brain anatomy, which we are only beginning to understand (Deacon, 1990).

The general evolutionary scenario outlined here entailed a shift to a new adaptive zone, not by an entire populous species but by a relatively small population of such a species that survived an environmental crisis. Other populations may have survived for a time with little change, in areas where environmental deterioration was less extensive. At least one fossil individual dated at about 1.6 Ma had a relatively small brain and more apelike proportions than individuals assigned unequivocally to early Homo (Leakey et al., 1989). In addition, two robust australopithecine species persisted well into Pleistocene time. The enormous molars and powerful jaw muscles of these forms endowed them with the ability to process a wider variety of plant foods than gracile forms, however, and this may have increased their chances for survival by reducing the need to migrate to new food supplies. Even these forms died out at about 1 Ma. This was approximately the time when glacial maxima and minima became more extreme (Stanley and Ruddiman, Chapter 7, this volume) and when carbon isotopes show that true savannas appeared (Cerling, 1992). Perhaps the increased severity of droughts during glacial maxima caused the extinction of the robust australopithecines.

There is evidence that Australopithecus africanus persisted to about 2.3 Ma (Delson, 1988), but we do not now know for sure that it survived beyond the origin of Homo at about 2.4 Ma. Thus, we cannot know for sure whether Homo emerged from the entire surviving population of the decimated ancestral australopithecine species or whether the ancestral species gave rise to Homo by the evolutionary divergence of just one of its populations and then survived for a time alongside it, though possibly in other geographic regions. Discovery of a temporal overlap within the poorly documented interval between 2.5 and 2.0 Ma would settle the issue in favor of evolutionary branching, as opposed to the bottlenecking of an entire species.

The mechanism of climatic forcing that I have described is compatible with either possibility, in that environmental deterioration must have been a complex process in time and space, and different populations were undoubtedly subjected to different patterns of environmental change. In any event, by mid-Pleistocene time, only the fully terrestrial genus Homo remained.

We tend to think of the environmental changes associated with the onset of the Plio-Pleistocene ice age as constituting a deterioration of habitats. Thus, it might seem a great irony that the origin of our genus, which we inevitably view as a positive event, was wrought by what, from a different perspective, has been widely viewed as an environmental crisis.

References

• Aiello, L., and C. Dean (1990). An Introduction to Human Evolutionary Anatomy , Academic Press, London, 596 pp.
• Bonnefille, R. (1976). Implications of pollen assemblage from the Koabi Fora Formation, East Rudolf, Kenya, Nature 264 , 487-491.
• Bonnefille, R. (1983). Evidence for a cooler and drier climate in the Ethiopian uplands towards 2.5 Myr ago, Nature 303 , 487-491.
• Brain, C. K. (1981). The Hunters or the Hunted? An Introduction to African Cave Taphonomy , University of Chicago Press, Chicago, 365 pp.
• Cartmill, M. (1974). Pads and claws in arboreal locomotion, in Primate Locomotion , F. A. Jenkins, editor. , ed., Academic Press, New York, pp.45-83.
• Cerling, T. (1992). Development of grasslands and savannahs in East Africa during the Neogene, Palaeogeography, Palaeoclimatology, Palaeoecology 97 , 241-247.
• Conrad, G. (1968). Evolution Continental Post-Hercynienne du Sahara Algerien , CNRS, Paris.
• Coque, R. (1962). La Tunisie Présaharienne , Amrand Colin, Paris, 476 pp.
• Count, E. W. (1947). Brain and body weight in man: Their antecedents in growth and evolution, Annals of the New York Academy of Sciences 46 , 993-1122.
• Deacon, T. W. (1990). Problems of ontogeny and phylogeny in brain-size evolution, Internat. J. Primatol. 11 , 237-282.
• Delson, E. (1988). Chronology of South African australopith site units, in Evolutionary History of the "Robust" Australopithecines , F. E. Grine, editor. , ed., Aldine de Gruyter, New York, pp.317-324.
• Devore, I., and S. L. Washburn (1963). Baboon ecology and human evolution, in African Ecology and Human Evolution , F. C. Howell, editor; and F. Bourliere, editor. , eds., Aldine, Chicago, pp.335-367.
• Harris, J. W. K. (1983). Cultural beginnings: Plio-Pleistocene archaeological occurrences from the Afar, Ethiopia, African Archaeol. Rev. 1 , 3-31.
• Hill, A. S., A. Deino, G. Curtiss, and R. Drake (1992). Earliest Homo , Nature 355 , 719-722. [PubMed: 1741057]
• Holt, A. B., D. B. Cheek, E. D. Mellits, and D. E. Hill (1975). Brain size and the relation of the primate to the nonprimate, in Fetal and Postnatal Cellular Growth , D. B. Cheek, editor. , ed., John Wiley & Sons, New York, pp.23-44.
• Jungers, W. L., and J. T. Stern (1983). Body proportions, skeletal allometry and locomotion in the Hadar hominids: A reply to Wolpoff, J. Human Evol. 12 , 673-684.
• Kay, R. F. (1985). Dental evidence for the diet of Australopithecus, Ann. Rev. Phys. Anthrop. 14 , 315-341.
• Kennedy, G. E. (1983). A morphometric and taxonomic assessment of a hominine femur from the lower member, Koobi Fora, Lake Turkana, Amer. J. Phys. Anthrop. 61 , 429-436. [PubMed: 6414307]
• Krogman, W. G. (1972). Child Growth , University of Michigan Press, Ann Arbor, 231 pp.
• Kruuk, H. (1972). The Spotted Hyena , University of Chicago Press, Chicago, 335 pp.
• Latimer, B., and C. O. Lovejoy (1990). Hallucal tarsometatarsal joint in Australopithecus afarensis , Amer. J. Phys. Anthrop. 82 , 125-133. [PubMed: 2360609]
• Latimer, B., J. C. Ohman, and C. O. Lovejoy (1987). Talocrural joint in African hominoids: Implications for Australopithecus afarensis , Amer. J. Phys. Anthrop. 74 , 155-175. [PubMed: 3122581]
• Leakey, M. D., and R. L. Hay (1979). Pliocene footprints in the Laetolil beds at Laetoli, northern Tanzania, Nature 202 , 7-9.
• Leakey, R. E. F., A. Walker, C. V. Ward, and H. M. Grausa (1989). A partial skeleton of a gracile hominid from the upper Burgi member of the Koobi Fora Formation, east Lake Turkana, Kenya, in Hominidae , G. Giacobini, editor. , ed., Jaca, Milan, pp.209-215.
• Marean, C. W. (1989). Saber-tooth cats and their relevance for early hominid diet and evolution, J. Human Evol. 18 , 559-582.
• McHenry, H. M. (1986). The first bipeds: A comparison of the A. afarensis and A. africanus postcranium and implications for the evolution of bipedalism, J. Human Evol. 15 , 177-191.
• McHenry, H. M. (1991). Sexual dimorphism in Australopithecus afarensis , J. Human Evol. 20 , 21-32.
• Rodman, P. S., and H. M. McHenry (1980). Bioenergetics and the origin of hominoid bipedalism, Amer. J. Phys. Anthropol. 63 , 371-378.
• Schaller, G. B. (1972). The Serengeti Lion , University of Chicago Press, Chicago, 480 pp.
• Servant, M., and S. Servant-Vildary (1980). L'environment quaternaire du bassin du Tchad, in The Sahara and the Nile , M. A. J. Williams, editor; and H. Faure, editor. , eds., Balkema, Rotterdam, pp.133-162.
• Skeat, W. W., and C. O. Blagden (1906). Pagan Races of the Malay Penninsula , MacMillan, New York, 724 pp.
• Stanley, S. M. (1992). An ecological theory for the origin of Homo , Paleobiology 18 , 237-257.
• Stern, J. T., and R. L. Susman (1983). The locomotory anatomy of Australopithecus afarensis , Amer. J. Phys. Anthrop. 60 , 279-317. [PubMed: 6405621]
• Susman, R. L., J. T. Stern, and W. L. Jungers (1984). Arboreality and bipedality in the Hadar hominids, Folia Primatol. 43 , 113-156. [PubMed: 6440837]
• Tague, R. G., and C. O. Lovejoy (1986). The obstetric pelvis of A. L. 288-1 (Lucy), J. Human Evol. 15 , 237-255.
• Vrba, E. S. (1974). Chronological and ecological implications of the fossil Bovidae at the Sterkfontein australopithecine site, Nature 250 , 19-23.
• Vrba, E. S. (1975). Some evidence of chronology and palaeoecology of Sterkfontein, Swartkrans and Kromdraai from the fossil Bovidae, Nature 254 , 301-304.
• Vrba, E. S. (1979). A new study of the scapula of Australopithecus africanus from Sterkfontein; Amer. J. Phys. Anthrop. 51 , 117-130.
• Vrba, E. S. (1980). The significance of bovid remains as indicators of environment and predation patterns, in Taphonomy and Paleoecology , A. K. Behrensmeyer, editor; and A. P. Hill, editor. , eds., University of Chicago Press, Chicago, pp.247-271.
• Vrba, E. S. (1985. a). African Bovidae: Evolutionary events since the Miocene, S. African J. Sci. 81 , 263-266.
• Vrba, E. S. (1985. b). Ecological and adaptive changes associated with early hominid evolution, in Ancestors: The Hard Evidence , E. Delson, editor. , ed., Alan R. Liss, New York, pp.63-71.
• Vrba, E. S. (1988). Late Pliocene climatic events and hominid evolution, in Evolutionary History of the "Robust" Australopithecines , F. E. Grine, editor. , ed., Aldine de Gruyter, New York, pp.405-426.
• Vrba, E. S., G. H. Denton, and M. L. Prentice (1989). Climatic influences on early hominid behavior, Ossa 14 , 127-156.
• Wesselman, H. B. (1985). Fossil micromammals as indicators of climatic change about 2.4 Myr ago in the Omo Valley, Ethiopia, S. African J. Sci. 81 , 260-261.
• Wood-Jones, F. (1900). Arboreal Man , E. Arnold, London, 230 PP.