464 Fossil humans

and 10–14% H with 2–5% S and the remainder oc-curring as N (< 1%) and O (< 1%). Natural gaswhich is mostly methane has about 75% C and about25% H.

Heating value. One important aspect of the fossilfuels is the heating value of the fuel, which is mea-sured as the amount of heat energy produced by thecomplete combustion of a unit quantity of the fuel.For solid fuels and usually for liquid fuels the heatingvalue is quoted for mass, whereas for gaseous fuelsthe heating value is quoted for volume. The heat-ing values are commonly expressed as British ther-mal units per pound (Btu/lb). In SI units the heat-ing values are quoted in megajoules per kilogram(MJ/kg). For gases, the heating values are expressedas Btu per cubic foot (Btu/ft3) or as megajoules percubic meter (MJ/m3). The table gives heating valuesof representative fuels. See ENERGY SOURCES; HEAT.

James G. SpeightBibliography. W. Francis and M. C. Peters, Fuels and

Fuel Technology, 2d ed., 1980; A. Markuszewski andB. D. Blaustein (eds.), Fossil Fuels Utilization: En-vironmental Concerns, 1986; R. A. Meyers (ed.),Coal Handbook, 1981; H. H. Schobert, Coal: TheEnergy Source of the Past and Future, 1987; J. G.Speight, The Chemistry and Technology of Coal,2d ed., 1994; J. G. Speight, The Chemistry and Tech-nology of Petroleum, 3d ed., 1999.

Fossil humansAll prehistoric skeletal remains of humans that arearcheologically earlier than the Neolithic (necessar-ily an imprecise limit), regardless of degree of min-eralization or fossilization of bone, and regardless ofwhether the remains may be classed as Homo sapi-ens sapiens (anatomically modern humans). In thissense, the term “humans” is used broadly to meanall primates related to living people since the lastcommon ancestor of people and African apes, thusall species currently included in the genera Homo,Australopithecus, Ardipithecus, and Paranthropus(and potentially others discussed below).

Discoveries began early in the nineteenth century,although their meaning and antiquity were not rec-ognized before the finding of the Neanderthal spec-imen in 1856, combined with Darwin’s demonstra-tion of the reality of evolution soon afterward. Fossilhuman remains have come principally from Europe,western Asia, China, Java, and Africa. Because of therather late entry of humans into the New World, allAmerican remains are of relatively recent origin andrecognizable as H. sapiens.

Dating fossils. The human lineage was oncethought to have come into existence only at the be-ginning of the Pleistocene, the geological time inter-val from about 1.8 million years ago to the present.It is now known to have been fully separate fromthe ancestry of the apes at least as far back as theEarly Pliocene or perhaps even the later Miocene(4–7 million years ago, or Ma). This view was widelyheld by anthropologists, on theoretical grounds, inthe early part of the twentieth century. It was aban-

doned in the 1940s because of lack of supporting evi-dence. Discoveries in Pliocene and Miocene depositssince then have led to a reappraisal of the evidence(Fig. 1).

Dating within the Pleistocene (and earlier) is ac-complished by the methods of conventional paleon-tology and geology, by association with human im-plements, by several chemical and physical tests forrelative age (such as the fluorine test for accumula-tion of this element in bone), and by geochronomet-ric methods which provide an age in years. The latterusually involve measuring the presence of radioac-tive substances such as radiocarbon and radiopotas-sium as well as other radioactivity-based methodssuch as electron spin resonance, thermolumines-cence, and fission-track dating. All of thesemethodsare broadly termed radiometric methods, as op-posed to approaches such as amino acid racemiza-tion or dendrochronology (tree-ring dating). See DAT-

ING METHODS; RADIOCARBON DATING.One of the most important methods of correla-

tion between dated geological sequences is paleo-magnetism, which does not itself provide a date be-cause it is basically a boolean system with two states(so-called normal and reversed polarity of geomag-netism) of which all occurrences are indistinguish-able. But if the pattern of reversals at any site canbe matched against the global paleomagnetic timescale and even approximately dated by means ofgeochronology or paleontology, a sequence of datesmay result. See PALEOMAGNETISM.

Prehuman ancestry. Humans are catarrhine pri-mates, part of a group including Old World monkeys,apes, and various extinct forms. Most evidence fromboth comparative morphology and molecular stud-ies of proteins shows that humans’ closest living rel-atives are the African apes: the chimpanzee and thegorilla. Less close is the Asian orangutan, and mostdistinctive of all apes are the gibbons. A classifica-tion which conforms to these relationships withinthe Hominoidea (apes, humans, and close extinctrelatives) recognizes the family Hylobatidae for thegibbons and Hominidae for humans and great apes.The latter family is divided into Ponginae (orangutansand extinct relatives) and Homininae. Many workers,however, continue to reject this view and place hu-mans in the Hominidae as contrasted with the pa-raphyletic (multi-origined) apes in Pongidae. A fewothers have suggested that in fact orangutans are theclosest living relatives of humans. No fossils of anymodern ape are known, other than Pleistocene gib-bon and orangutan teeth from southern Asia, and afew Kenyan teeth identified as chimpanzee whichwere recognized in 2005. See FOSSIL APES; FOSSIL

PRIMATES; MOLECULAR ANTHROPOLOGY; MONKEY;PRIMATES.

The oldest certain representatives of the Catar-rhini are fossils from the Fayum beds of north-ern Egypt dated around 34 Ma. The best knownis Propliopithecus (=Aegyptopithecus) zeuxis, aspecies near the common ancestry of apes, hu-mans, and Old World monkeys. Hominoids origi-nated in Africa, probably during the Late Oligoceneand Early Miocene (26–17 Ma), as suggested by the

Fossil humans 465

known time rangeof form named

unknown time range

Key:

ancestry

Propliopithecus(Aegyptopithecus)

to Old World monkeys

Proconsulidae

to orangutan

Dryopithecinae

GraecopithecusSamburupithecus

to gibbons

Ankarapithecus,Sivapithecus,

Gigantopithecus

LATE

MIDDLE

EARLY

Gigantopithecus

Homo floresiensis

Homoerectus

Australopithecusafarensis

?

?

Neanderthals

Homorhodesiensis

Homo sapiens(modern humans)

Homo habilisHomo rudolfensis

Paranthropusrobustus

Australopithecusafricanus

Paranthropusboisei

Paranthropusaethiopicus

chimp

Australopithecusanamensis

Ardipithecusramidus

?

to African apes

? Sahelanthropus

? Ardipithecus kadabba

0

.05

.1

.25

.5

1.0

2.0

3.0

5.0

10

15

20

25

30

35

PLEI

STO

CEN

EPL

IOCE

NE

MIO

CEN

EO

LIG

OCE

NE

H. mauritanicus

Milli

ons

of y

ears

ago

H. neanderthalensis

Kenyapithecinae

Fig. 1. Human phylogeny from the Oligocene to the present time, showing the skulls of the major known fossil relatives andpossible ancestors of modern humans.

466 Fossil humans

genera Kamoyapithecus and Proconsul, placed inthe family Proconsulidae. However, some scholarshave questioned the view that these forms are al-ready hominoids, proposing instead that they wereconservative catarrhines that predate the evolution-ary split between hominoids and Old World Mon-keys. Between 20 and 15 Ma, there are several morederived (“advanced”) African hominoid genera (forexample, Morotopithecus, Afropithecus, and Kenya-pithecus) and one Eurasian form (Griphopithecus).If these are more closely related to living great apesthan are gibbons, as has been proposed in the past,they may represent early members of the Hominidae.However, if they are conservative (“primitive,” ple-siomorphic) by comparison with the gibbons, thenthey should not be included in the hominid family.

In the latter case, several workers have arguedthat the origin of modern ape (and human) lineagesmust have occurred in Eurasia, where a number ofderived genera occur between 14 and 8 Ma. Theseinclude Dryopithecus and Pierolapithecus (discov-ered in 2004), which were probably close to thecommon ancestor of all great apes and humans;Ankarapithecus and Sivapithecus, which appear torepresent two early stages in the orangutan lineage(Ponginae); and Graecopithecus (Fig. 2; also termedOuranopithecus), which has been suggested as amember of Homininae, that is on the lineage lead-ing toward African apes and humans. Another earlyhominine is Samburupithecus, known only froman upper jaw fragment found in the Samburu Hillsof Kenya and (like Graecopithecus) dated to about9.5 Ma. A possible alternative interpretation is that aform such as Morotopithecus, with relatively derivedbut poorly known postcranial elements, might havegiven rise both to the early Eurasian hominids and to

Fig. 2. Face (cast) of Graecopithecus freybergi fromXirochori, Greece, about 9.5 million years old. (Courtesy ofand c© E. Delson; photo by L. Meeker)

an as yet unknown African line which culminated inSamburupithecus. In either case, it is not until thelatest Miocene (7–5 Ma) that we see the first possi-ble fossil representatives of the Hominini (hominins),the group including modern humans and all our rel-atives since the last common ancestor shared withchimpanzees. Molecular anthropological studies ofDNA sequences and other lines of evidence have sug-gested that chimpanzees are more closely related tohumans than either is to gorillas. This would appearcounterintuitive, given the morphological similaritybetween the two African apes, both of which arealso apparently derived in their knuckle-walking lo-comotion. But similarities may be merely holdoversfrom the common ancestor of all three forms, whilechimps and humans shared an intermediate commonancestor after the gorilla lineage split away. Paleonto-logical evidence is sparse in Africa between 12 and5 Ma, but it is possible that Samburupithecus (orGraecopithecus) represents a species near the baseof that three-way split or already on the gorilla line.Their ages of 10–9 Ma fit reasonably well with the“molecular clock” estimates of 10–7 Ma for that di-vergence. The human–chimp split is molecularly es-timated at 8–5 Ma or even less, but the human fos-sil record refutes a date younger than perhaps 5 Maat the minimum. See FOSSIL APES; FOSSIL PRIMATES;MOLECULAR ANTHROPOLOGY; MONKEY; PRIMATES.

Identification of early hominins. When paleoanthro-pologists find early fossils that might be on thehuman lineage, they must identify them as such bylocating bony features which distinguish the earliesthominins from their ape relatives or ancestors.

In order to find such features in fossils, we mustbegin by determining what makes modern humansdifferent from living apes. We can then discern thesequence in which such features appeared in thehuman fossil record. However, not all changes hap-pened at the same time. This pattern is known asmosaic evolution, referring to mosaic pictures madeup of numerous small parts which fit together tomake a complete image.

One major difference involves the mode of lo-comotion used by hominins versus apes. Homininsare habitually bipedal animals (walking on two legsrather than four), whereas our ape ancestor likelyincluded quadrupedal, suspensory and/or knuckle-walking behaviors in its locomotor repertoire. Thehuman skeleton has been modified in many ways toaccommodate this change. For example, homininshave a foot with a longer big toe that is in linewith the other toes, rather than a grasping ape-like foot with a divergent big toe. The foot hasalso developed two arches, one longitudinal andone transverse (along the long and short axes ofthe foot, respectively) that act as shock absorberswhen walking bipedally. While the spine of apesis C-shaped, hominins evolved thoracic and lumbarcurvatures, resulting in an S-shaped spine that actsto balance the torso over the pelvis and legs. Thisspine also connects to the skull more vertically (andcentrally) underneath rather than toward the back ofthe skull (as in apes), which can be seen by the more

Fossil humans 467

forward placement of the foramen magnum (open-ing where the spinal cord enters the skull). The pelvisand pelvic muscles have also gone through a majorreorganization: the pelvis is less tall and more bowl-shaped, and the gluteus maximus muscle is more de-veloped to help extend the leg backward in walking.The femur (bone of the thigh) is more angled in at theknee in order to center the feet beneath the body, andthe knee is capable of straightening out completely.Modern humans also have quite long legs and onlymoderately long arms compared to body size. All ofthese features aid in more stable and efficient bipedallocomotion in hominins than apes. The fossil recordsuggests that relatively modern-looking spinal cur-vature, pelvic shape, and femoral angling appearedearly, while foot arching, perhaps big toe position,and especially fully modern limb proportions did notevolve until later in the hominin lineage. The humanskull has also undergone changes not directly relatedto bipedal locomotion, such as an increase in brainsize (relative to body size) and changes in the teeth,including a decrease in canine size. While some earlyhominins have relatively larger brains than chim-panzees of similar body weight, major brain expan-sion probably did not occur until the earliest memberof our own genus Homo evolved.

The next question then is: When in the fossilrecord do we first see evidence of these changes?Recently discovered fossils from Chad (termed Sa-helanthropus; Fig. 3) may include the oldest ho-minin cranium with smaller canines and a forwardlypositioned foramen magnum (which may in turnsuggest some bipedalism). Limb bone fossils fromKenya (named Orrorin) are said to have a number ofcharacteristics consistent with bipedalism, thoughsome debate surrounds these claims. Both sets offossils probably date from 7–6 Ma, making themthe oldest possible hominins in existence. Youngerfinds (∼4.4 Ma) from Ethiopia (Ardipithecus) havemuch smaller canines and may also have a foramenmagnum that is positioned more forward than in

Fig. 3. Cranium of Sahelanthropus tchadensis from deposits ca. 6–7 million years old in Chad, original fossil (left) and virtualreconstruction. (Courtesy of and c© M. P. F. T.)

apes. Canine reduction may have begun even be-fore hominins appeared, as seen in the fossil apeGraecopithecus. Further decreases in size eventu-ally led to the very small canines present in mod-ern humans. It has been suggested that smaller ca-nines (with no significant difference between thoseof males and females) reflect a different social or-ganization, because large canines are associated inapes (and monkeys) with a high degree of com-petition among males for mates. Fossil australo-piths (see below) show clear evidence of bipedalismboth in their skeletons and in fossilized footprintsfound at the site of Laetoli, Tanzania. These homininsstill display many apelike features in their skull, show-ing that habitual bipedalism evolved before modernskull anatomy.

Miocene Hominini. In 2002, French and Chadianresearchers led by Michel Brunet reported the dis-covery of new fossils from Chad in Central Africa.Based on the associated mammalian fossils, the siteof Toros-Menalla was estimated to date about 7–6Ma. Previously, almost all early human remains hadbeen found in eastern or southern Africa, althougha few fragments from Chad dating to about 3.5 Mahad been known since the 1990s. The new fos-sils (named Sahelanthropus tchadensis) included afairly complete though crushed cranium (the skullwithout the lower jaw) of what might be the oldesthominin ever found. Recent computer-based analy-sis has permitted a “virtual” reconstruction of whatthe cranium might have looked like before it wascrushed (Fig. 3). Although the brain size is no biggerthan that of a similarly sized ape, its teeth are morelike a hominins, especially in having an apparently re-duced canine. The relatively anterior position of theforamen magnum suggested to the discoverers thatthis species may have been at least partly bipedal, butothers have questioned this view. Sahelanthropus islikely near to the chimpanzee–human split, based onboth its morphology and the molecular divergencedate discussed above.

468 Fossil humans

In late 1999, Paris-based Martin Pickford andBrigitte Senut working in Kenya discovered a smallcollection of fragmentary fossils, mostly limb ele-ments, which they also claimed as the earliest knownhominin. The fossils were first termed “MilleniumMan,” as they were announced to the public justbefore 2000, but they have been formally namedOrrorin tugenensis. These remains are better datedthan those of Sahelanthropus, to about 6 Ma, by ra-diometric techniques. It has been claimed that the fe-mur shows features indicating bipedalism, but mostresearchers have questioned this interpretation, andsome have suggested that Orrorin is actually thelong-sought fossil of an early African ape ancestor.Unfortunately, dental and cranial remains are toofragmentary to be certain either way.

Still younger fossils are known from sites in theMiddle Awash Valley in Ethiopia. Recovered between1997 and 2004 and first described in 2001, thesewere named Ardipithecus kadabba in 2005. Thefossils range in age between 5.8 and 5.2 Ma, at thevery end of the Miocene epoch. Most are fragmen-tary jaws, but a single toe bone has been claimedto indicate bipedalism, although (as usual), some re-searchers have questioned this interpretation.

Pliocene Hominins: the australopiths. Pliocene hu-mans have been grouped in various ways, but it nowseems that four main types can be distinguished.Three of these, dating from roughly 4.5–1 Ma, havepreviously been assigned to the genus Australopithe-cus and can informally be termed australopiths (herethey are placed in the genera Ardipithecus, Aus-tralopithecus, and Paranthropus). The fourth groupincludes early species of Homo, beginning about2.5 Ma. The australopiths have previously been di-vided into gracile and robust varieties, but the for-mer term really is not accurate and is not used here.All australopith species appear to share a number ofbasic characteristics (see above) distinguishing themfrom living and fossil apes and also from later hu-mans, although clearly linking them to the latter. Thespecies of australopith also broadly share a smallerbody size than modern humans with surprisingly lit-tle difference among the known forms. Estimates ofbody size based on analysis of weight-bearing jointsurfaces and bone lengths suggest average weightsaround 130 lb (59 kg), with females often under 77 lb(35 kg) and males over 155 lb (70 kg). This quitehigh sexual dimorphism is also typical of australo-piths, as are back teeth that are large for the esti-mated body size, although the actual sizes and toothproportions are among the features distinguishingthe several species from one another. In light of thediversity now recognized, many researchers accept adivision into the genera Australopithecus and Paran-thropus, as well as Ardipithecus, but others continueto recognize only one or two of these genera. SeeAUSTRALOPITHECINE.

The fossils of these early humans were first foundin South Africa in 1924, but the most recent majordiscoveries and the best evidence of their age comefrom East Africa. The South African fossils come fromsix main sites, which are the remains of ancient cave

systems. The australopiths did not live in caves, buttheir carcasses may have been washed or droppedinto crevices near the ground surface by leopards orother carnivores. In two sites especially, many fossilsare known from relatively short spans of time (per-haps less than 200 thousand years at each), but theevidence for dating is not definite.

In 1959 Mary Leakey and L. S. B. Leakey discov-ered a nearly complete australopith skull at Oldu-vai Gorge, Tanzania, and colleagues dated it atabout 1.75 Ma, far older than previously thought.Since then, American and joint American-French andAmerican-Ethiopian expeditions have found hun-dreds of human fossils in the Omo and Afar areas ofEthiopia, while a Kenyan team has worked aroundthe shores of Lake Turkana (formerly Lake Rudolf),Kenya, and various studies have continued at Oldu-vai and nearby Laetoli. These regions have yieldedsmaller numbers of specimens at many separate sub-sites, but the age of each site can usually be estimatedclosely by potassium-argon and paleomagnetic dat-ing. Also, many specimens are more complete andshow less distortion than their contemporaries fromSouth Africa. Until 1993, no definite examples of Aus-tralopithecus or Paranthropus were known outsidethese areas, although some claims have been made.The recovery of australopith specimens from Chad(and of Paranthropus and early Homo from Malawi)demonstrates that new productive regions remainto be explored. It is not clear whether australopithsmade stone tools, although they were probably col-lectors of plant foods, using unpreserved woodensticks and skin bags, and perhaps scavengers of smallgame animals.

Ardipithecus: still a question. The oldest known likelyhuman species, Ardipithecus ramidus, is knownfrom a small group of fossils found at the Aramis local-ity in the Middle Awash Valley, Ethiopia. (Additionalremains of the species have been reported from thenearby Gona locality as well.) The Aramis fossils havebeen recovered since 1993 from rocks just abovea layer dated to 4.4 Ma, which also yielded animaland plant fossils suggesting a relatively wooded en-vironment. The remains of Ard. ramidus includeteeth, part of a skull base, and the almost completelong bones of a forelimb, as well as many pieces ofa fragmented adult skeleton, including parts of al-most all regions of the body. As of early 2006, thisskeleton had not yet been formally described or il-lustrated, but the skull is said to be crushed flat andfragmented into hundreds of tiny pieces which havenot yet been reconstructed. The elements describedso far document a mosaic pattern combining fea-tures similar to those of younger humans (for exam-ple, the anterior position of the foramen magnumand nearby structures, implying upright posture andpresumably bipedalism; and partial reduction of thecanines) with others reflecting retention of apelikeconditions (for example, thin enamel covering onmolars and incisors; lack of a second cusp, or meta-conid, on the lower anterior premolar; relativelylarge size of canines; and apelike shape of the loweranterior deciduous premolar or “milk molar”). It has

Fossil humans 469

been suggested that their presence in a wooded en-vironment implies that australopiths may have differ-entiated from apelike ancestors in forests rather thanopen savannahs. But an alternative view suggests thatthis species represents a “failed” human lineage, onewhich returned to the forest and secondarily devel-oped thinner enamel convergently with African apes.The remaining bony features of Ard. ramidus are allreasonably interpreted as ancestral conditions, to beexpected in an ancient human ancestor. Analysis ofthe partial skeleton should permit determination ofthe locomotor abilities of this species. Other speci-mens from Kenya might be referable to this speciesbut are too fragmentary to be sure yet.

Earliest definite humans: Australopithecus. The first habit-ual bipeds, Australopithecus, appear in the fossilrecord in quantity around 4.2 Ma, during the EarlyPliocene. The earliest representatives of Australo-pithecus, named Aus. anamensis, have been foundsince 1994 at sites in the southern Lake Turkana re-gion of Kenya. From Kanapoi, in a layer dated to4.2–4.1 Ma, come several jaws that differ in shapefrom those of other Australopithecus species. A par-tial ankle joint from a younger layer confirms uprightwalking around 4 Ma. It seems likely that Aus. ana-mensis may be close to the ancestry of later speciesand a possible descendant of Ard. ramidus.

Fossils from sites in Ethiopia and Tanzania revealfar more details about a still younger species, Aus.afarensis (Fig. 4). The most complete material isknown from the Ethiopian site of Hadar, about 50 km(30 mi) north of Aramis, where deposits yielded fos-sils dating between 3.4 and 2.9 Ma. In 1974 a partialskeleton was found and identified as a female by itspelvic bones (and small size compared to other fos-sils) and nicknamed Lucy. This individual would havestood only 3.5 ft (106 cm) tall and weighed perhaps65 lb (30 kg). The leg bones of this skeleton indicatethat Lucy’s legs would have been rather short for amodern human of comparable body weight, but theywere in proportion for her estimated stature; thisobservation supports other evidence that most aus-tralopiths were more robustly built than modern hu-mans. Larger males, perhaps 130–175 lb (59–80 kg)in weight, are known from more fragmentary re-mains, although a nearly complete skull (not knownfor Lucy) was found in the 1990s and analyzed in

Fig. 4. Cranium of male Australopithecus afarensis fromHadar, Ethiopia, dated to about 3.1 million years ago.(Courtesy of and c© Institute of Human Origins)

detail in 2004 (Fig. 4). Additional specimens, includ-ing partial skulls, continue to be found at Hadar. Inaddition, mandibles and postcranial elements fromMaka (across the Awash River from Aramis) and Lae-toli (Tanzania) dating around 3.4–3.7 Ma confirm thesexually dimorphic but taxonomically unified natureof the species, while a frontal bone from Belohdelie(3.8 Ma, near Maka) and teeth from the Lake TurkanaBasin (3–2.7 Ma) define its temporal range. A frag-mentary lower jaw from Chad is estimated to dateto a similar time interval by comparing the animalfossils associated with it to those from East Africa; ithas been given a new species name but could be awestern representative of Aus. afarensis.

Lucy’s pelvis and leg bones, as well as remarkablypreserved footprints from Laetoli, clearly demon-strate that upright bipedal walking was well devel-oped by 3.6 Ma, along with a brain somewhat largerthan in modern apes of similar body size. Brain sizein Aus. afarensis might have been between 350 and450 ml, in a body weighing some 65–154 lb (30–70 kg) as compared to 365 ml in a 100-lb (45-kg)chimpanzee, 500 ml in a 300-lb (135-kg) gorilla, or1400 ml in a 150-lb (67-kg) living human. Argumentsas to the priority of brain expansion or walking abil-ity in human evolution thus have yet to be fullyresolved.

Australopithecus afarensis combines both ofthese advanced, human characteristics with numer-ous other features reminiscent of later Miocene ho-minids and modern apes. The lower face is ratherprojecting (prognathic), the canines project slightlybeyond the level of the neighboring teeth, the ante-rior lower premolar sometimes has only one majorcusp, the tooth row is elongated and nearly parallel-sided, and the forehead is low and retreating. How-ever, the mastoid region (below and behind the bonyear opening in the skull) projects inferiorly morethan in either living apes or modern humans. More-over, the long curved fingers and toes support theidea that it often climbed trees, perhaps to sleep,although there is still some controversy over the de-gree of posible arboreal behavior in this species.

The geologically youngest of the early australo-piths is actually the first to have been recognized.Australopithecus africanus was named by R. Dartin 1925 on the basis of a juvenile specimen fromTaung (South Africa). It appears to have lived from3 to 2.3 Ma, but so far it is known only from foursites in South Africa; by far the most numerous spec-imens come from the middle levels (Member 4) ofSterkfontein cave, dated about 2.8–2.6 Ma (Fig. 5).It may have stood 4–5 ft (120–150 cm) tall, weighed65–150 lb (30–68 kg), and had an average brain sizeof some 450 ml. The skull seems more lightly builtthan in Aus. afarensis, with a rounded vault but moreprojecting face. The teeth are more humanlike aswell, especially in the presence of two cusps onall anterior lower premolars (like all younger ho-minins) and less projecting canines. Although theabsolute tooth size of Aus. africanus is nearly equalto that of small gorillas, the proportions are human,with a smooth decrease in size from molars through

470 Fossil humans

Fig. 5. Cranium of adult (female?) Australopithecusafricanus from Sterkfontein, South Africa, about 2.5 millionyears old. (Courtesy of I. Tattersall)

incisors. Its postcranial skeleton appears quite simi-lar to that of Australopithecus afarensis, with a fullybipedal (if by no means modern) locomotor adap-tation. Thus, it was perhaps little changed from itsputative ancestor, but those few changes are roughlyin the direction of later humans. Some workers havesuggested that it might lie close to the ancestry ofthe genus Homo, but pending clearer evidence forits phyletic position, that step is not taken here.

In 1995 researchers described bones of a par-tial foot from the lower levels (Member 2) ofSterkfontein, perhaps contemporaneous with Aus.afarensis. These bones were interpreted as indicat-ing a foot partially adapted to grasping and perhapstree climbing rather than only to bipedal walking. Inlate 1998 the find of additional parts of the sameskeleton was reported, including both legs, somearm bones, and what appears to be most of the skull.All of these elements must be fully removed fromthe encasing rock and reconstructed before they canbe accurately analyzed, but much more of the skele-ton may be preserved, in which case it might sur-pass Lucy in its completeness. Dates between 4 and3 Ma have been suggested by different techniques(and some scholars even think it might be youngerthan 2.5 Ma), and it has not yet been determined if itshould be identified as Aus. africanus, Aus. afaren-sis, or a new species of australopith.

In 1999, Ethiopian and American researchers de-scribed Aus. garhi from deposits in the Awash valleyof Ethiopia, dated to about 2.5 Ma. A partial face anduncertainly associated limb bones were said to repre-sent a new species close to the origin of Homo, butother workers have suggested alternative interpre-tations, such as that these fossils represented a latepopulation of Aus. afarensis or female individuals ofPar. aethiopicus. Stone tools and animal bones withstone tool cut marks were also found in the vicin-ity, raising the possibility that australopiths may havemanufactured stone tools. On the other hand, earlymembers of our own genus, Homo, first appear inthe fossil record at other sites around this time pe-riod and may have been responsible for making andusing these tools.

Robust varieties. Until 1986, robust australopithswere known from two forms, usually accepted as dis-

tinct species and increasingly given generic status asParanthropus: P. robustus in South Africa and themore extreme P. boisei in East Africa. These specieslived between about 2.3 and 1.4 Ma (the age ofP. robustus is known with less certainty, perhaps1.9–1.6 Ma) and are distinguished from the other aus-tralopiths by their larger size and craniodental spe-cializations. They may have been 4 ft, 6 in. to 5 ft,9 in. (135–175 cm) tall, weighed 80 to 190 lb (36 to86 kg), had a somewhat heavy and muscular bodybuild, and a brain size of about 525 ml. The skull isrobust, with deep cheekbones and thick lower jawand often a slightly raised sagittal (midline) crest inthe middle part of the skull roof from back to front.These features indicate strong chewing muscles andperhaps a diet of tough foods. The teeth themselvesare distinctive: the back teeth (molars and premo-lars) are large to huge; the front teeth (incisors andcanines) are quite small and run nearly straight acrossthe front of the mouth. This difference from otherhominins, including apes (which generally have largefront teeth and small back teeth), combined with alow forehead and a concave, nearly upright face, fur-ther suggests adaptation to powerful chewing. Theanterior teeth were probably used as much for grind-ing as for cutting. The scrappy postcranial elementsreveal no major adaptive differences from those ofAustralopithecus.

An even more ancient species, (P. aethiopicus),presents a combination of many of these Paranthro-pus features plus others that appear to be holdoversfrom an Aus. afarensis–like ancestry. Here, large mo-lars lie behind sockets for rather large incisors andcanines; a concave upper face with low foreheadsits above a projecting snout; and the sagittal crestextends back to meet the large nuchal (neck mus-cle) crest (Fig. 6). The brain size was small, near400 ml. The combination indicates that this speciesemphasized both large front and back teeth, occupy-ing an evolutionary position intermediate betweenAus. afarensis and the two previously known robustaustralopiths; its large mastoid area also links it toAus. afarensis. It is also intermediate in time, rang-ing from 2.7 to 2.3 Ma.

Australopith relationships. The widely accepted viewbefore 1978 was that Aus. africanus represented thecommon ancestor of the robust forms and Homo.The discovery of Aus. afarensis led to placement ofthis species as the basal hominin, with several al-ternative views of its descendants. In the 1990s thenewly recovered Aus. anamensis was usually rankedas a still older common ancestor, with Ard. ramidusheld off to the side, in suspense. Early analyses ofP. aethiopicus considered it as close to the com-mon ancestor of P. boisei and P. robustus, but somelater studies argued that the three robust specieswere “only” linked by their common possession of aheavily built chewing apparatus, which might haveevolved convergently in two or more disparate lin-eages. These studies proposed (as had earlier work-ers on other grounds) that Aus. africanus might havebeen ancestral to P. robustus in South Africa or thatP. aethiopicus might have been an early experimentunrelated to later robust species. Such views seem

Fossil humans 471

Fig. 6. Cranium of adult (male?) Paranthropus aethiopicusfrom West Turkana, Kenya, dated to about 2.5 million yearsago. (Courtesy of A. Walker)

poorly founded, for if the three robust species are infact linked by the adaptive complex of “heavy chew-ing” composed of a number of closely similar ele-ments in each form, it is far more likely that theydeveloped from a single common ancestor than thatthey were convergently comparable but unrelated.Here the robust clade or lineage is recognized as anevolutionary unit, the genus Paranthropus, whichwas thus long-lived but seems to have had no laterdescendants.

No known australopith clearly shows features thatlink it to later human species of the genus Homo, butmany workers have suggested that Aus. africanusmight represent the closest approach to such an an-cestry yet recovered. However, some have thoughtthat several shared features of the skull point to acommon ancestry of Homo and Paranthropus tothe exclusion of Aus. africanus (see also below). IfAus. afarensis is eventually determined to have been(close to) the common ancestor of both Paranthro-pus and Homo (plus Aus. africanus?), then it willprobably be wise to provide a new genus name forafarensis (and also anamensis), but for the momentall three species are retained in the genus Australo-pithecus. See AUSTRALOPITHECINE.

The rise of Homo. The only other genus of the Ho-minini is Homo, true humans, into which all laterforms are placed. To summarize in advance, in thisessay we suggest that after the brief occurrence oftwo or three early species of Homo in Africa, H. erec-tus evolved about 2 Ma and soon spread into Eurasiaeast of continental Europe; its far-flung populationswere connected by periodic population movementsspreading genetic modifications and keeping thespecies relatively unified. Homo erectus persisted,at least in eastern Asia, until 0.2 Ma or later, but far-ther west around 1 Ma, a geographically restrictedspecies arose near the Mediterranean, with possi-ble representatives in southern Europe and northernAfrica. This new species in turn gave rise to at leasttwo major lineages: one in Europe led to the Nean-derthals, while a second in Africa continued withlittle change for half a million years. By about 200 Ka

(Ka = thousand years ago) our own species Homosapiens arose in Africa and slowly spread across thatcontinent and into Eurasia, where it eventually re-placed Neanderthals and all other human forms.

Early Homo. The identification of the earliest spec-imens of Homo is a subject of debate among pa-leoanthropologists. In the late 1970s the scientificpendulum had swung back to an idea proposed onless secure grounds by L. S. B. Leakey and colleaguesin 1964. They named the species H. habilis, basedon several finds from Olduvai. Especially significantwas the discovery of the remains of a juvenile’s lowerjaw, with teeth much like those of Aus. africanus,and its partial skull, with an estimated cranial capac-ity of about 685 ml, dated about 1.8 Ma. After muchargument over the “reality” and distinctiveness of thenew species, it was made clear from additional findsat Olduvai, Lake Turkana, and probably a youngerlevel at Sterkfontein (Fig. 7) that a relatively small-brained (510–700 ml) and small-toothed Homo waspresent during 2.0–1.5 Ma. This was thought to beyounger than Australopithecus, older than (most)H. erectus, and contemporaneous with P. boisei. Apartial skeleton discovered at Olduvai in 1986 hassimilar teeth and an estimated body size comparableto Lucy; this small size was reported as remarkablebut should have been expected, given the similarityin skull size to those of the smaller australopiths.

Several fossils, especially from the Lake Turkanaregion, appeared to represent a different “morph”or structural pattern. These were typified by skullKNM-ER 1470 (its catalog number in the Kenya Na-tional Museum) which has a relatively flat but pro-truding face, a brain size of about 750 ml, a highrounded vault, and probably large teeth (the crownsare broken off). This (and a few more fragmentaryspecimens) were at first suggested to represent maleindividuals, while the smaller individuals were con-sidered females of the same species. Leg bones sug-gesting a size of 4 ft, 9 in. (155 cm) and 110 lb(50 kg) were originally thought to go with thesecranial parts, but it later became clear that earlyH. erectus overlapped in time with the more archaicfossils and had comparable leg bones. Unassociated

Fig. 7. Partial cranium of probable Homo habilis fromSterkfontein, South Africa, about 1.9 million years old.(Courtesy of the late A. R. Hughes)

472 Fossil humans

postcranial elements have had to be set aside pend-ing the recovery of more bones clearly linked to iden-tifiable crania.

One species or two?. These two sets of early humanfossils pose an active problem in paleoanthropol-ogy: either H. habilis had greater sexual dimorphism(especially in brain size) than any known primate, ortwo (rather homogeneous) species were masquerad-ing under a single name, with the smaller set (thesupposed females) most similar in facial morphologyto both Aus. africanus and later Homo. There are atleast two opposing solutions to this problem. Oneview holds that all the known specimens from Oldu-vai, the Turkana Basin, and South Africa represent thesingle species H. habilis, a larger-brained and biggercreature than Aus. africanus, but one with essen-tially the same dental apparatus—another example ofmosaic evolution. It would have had as much sexualdimorphism as did Aus. afarensis, perhaps at leastas much as in modern gorillas or orangutans, in bothform and size of teeth, face, and brain.

Another suggestion is that almost all of the Oldu-vai fossils, the smaller Turkana region specimensand some from South Africa, are recognized as H.habilis, while the 1470 specimen and other larger(non-Paranthropus) individuals from Turkana areconsidered as H. rudolfensis. This two-species viewis accepted here. In 1992 a partial mandible wasfound at Uraha in Malawi that is extremely simi-lar to a Turkana H. rudolfensis specimen. Not onlyis Malawi geographically intermediate between theknown eastern and southern African site regions,but preliminary age estimates based on associatedfossil mammals place the find at about 2.5–2.3 Ma,making it one of the oldest representatives of thegenus Homo. A temporal bone (the lower middlepart of the cranium, where the mandible hingesand the ear is housed) of similar age from centralKenya might also represent this species, and someauthorities have even suggested that a few Olduvaispecimens belong here as well. One new fossil (thelower part of a face with teeth) described in 2003from the western part of Olduvai shares certain sim-ilarities with both the 1470 fossil and the Olduvailower jaw first included in H. habilis. The describersof this new specimen suggested that the nameH. habilis would thus best be applied to the larger-sized group (including what had previously beencalled H. rudolfensis), requiring a new name for theremaining smaller fossils; the jury is still out. To fur-ther complicate matters, some researchers have sug-gested that 1470 and its closest relatives might belinked to another Kenyan cranium found in 2001 anddated close to 3.5 Ma. Termed Kenyanthropus platy-ops (“flat-faced Kenya man”), this fossil is somewhat“squashed” and its surface bone broken into manysmall, slightly separated pieces. It might represent anew lineage of Pliocene hominins; that lineage mightinclude 1470, renamed as K. rudolfensis; or the newfossil might just be a badly damaged representativeof a known species, such as A. afarensis. For themoment, it seems wisest to accept just two namedearly species of Homo, with other options awaitingstronger supporting evidence.

Although australopiths were habitually bipedal,limb bones questionably attributed to H. rudolfen-sis suggest a wider pelvis, perhaps to permit thebirth of larger-headed (because larger-brained) in-fants, which resulted also in the leg bones themselvesappearing more modern. It is this combination of ap-parently significant change in the two major humanadaptations of locomotion and intelligence that leadsmost authors to classify these species as Homo, butsome have placed both of them in Australopithecus.Those researchers take the position that the anatomyof H. habilis and H. rudolfensis suggests their “adap-tive strategies” were more like that of Australopithe-cus than other Homo species. The problem withthis argument is that most other studies support astronger evolutionary relationship between H. ha-bilis, H. rudolfensis, and the other members of thegenus Homo, including H. erectus, Neanderthals andmodern humans, than with Australopithecus. Whilethe former idea has not taken hold widely in the pa-leoanthropological community, it does suggest thata fundamental shift in adaptive strategies took placeafter these early Homo species.

It is not clear which species of early Homo mighthave been ancestral in turn to H. erectus. Both earlyforms have been found in the 2.0–1.6 Ma time rangein the Turkana Basin and probably Olduvai, wherethey overlapped with P. boisei and H. erectus. Butneither H. habilis nor H. rudolfensis as known aremorphologically very close to early H. erectus. Theformer species differs not only in small brain andbody size but also in limb proportions, although itsfacial morphology is potentially acceptable in an an-cestor of H. erectus. The brain of H. rudolfensis islarge, as might be expected in an ancestor of H. erec-tus, but its teeth are large and distinctive, its face isunique, and its limb bones have not been satisfac-torily identified. One wonders if there might not bean as yet undiscovered early species of Homo whichcombined smaller teeth and face with a larger brain.Perhaps the upper jaw from Makaamitalu (foundin 1994 high in the Hadar sequence) dated about2.35 Ma might be a step toward fulfilling this predic-tion.

Early Homo technology. Evidence of archeological ac-tivity has been found with or near some of theseearly Homo specimens, mostly flakes and choppersor “pebble tools” of the Oldowan stone-tool indus-try (see table), and also the remains of small an-imal prey. It is not possible to tell which typesof humans made the tools, but at present the old-est evidence is from Ethiopia, in the Gona region(near Hadar and Aramis) and the Omo valley, about2.5 Ma. The close time and space concordance ofOldowan tools and early Homo suggests that atleast one of these species was the actual toolmaker;whichever it was may have been an active hunterof small game, collector of plant and insectfood items, and scavenger of larger mammals.See PALEOLITHIC; PREHISTORIC TECHNOLOGY.

Homo erectus. While H. habilis and H. rudolfen-sis apparently were short-lived and relatively rareAfrican species, their likely successor, H. erectus,was common, widespread, and long-surviving. The

Fossil humans 473

Old World archeological variants before 10,000 years ago

Technological Paleolithicmode Description subdivisions Selected regional industries∗

4 Blade cores, to produce scrapers,points, burins; bone harpoons, art

Late (Upper) Paleolithic Later Stone Age [sub-Saharan Africa,40 to < 5 Ka (mode 5, microlithic)]

Upper Paleolithic of Europe, includingPerigordian Solutrean, Magdalenian

Aurignacian (Europe, SW Asia?, > 40to 27 Ka)

3 Widespread use of prepared cores toobtain variety of flake forms, usedas scrapers and projectile points

Middle Paleolithic Chatelperronian (western Europe,36–30 Ka)

Aterian (northern Africa, 100 to 30 Ka)Mousterian (many variants; western

Eurasia and northwestern Africa,200 to 35 Ka)

Middle Stone Age (many variants;sub-Saharan Africa and southernAsia, 250 to 40 Ka)

2 Large bifaces (handaxes andcleavers), also simple cores andretouched flakes

Early (Lower) Paleolithic In late stages, use of prepared coresand wooden throwing spears

Acheulean (Africa and southwesternAsia, 1.5 to <0.2 Ma; eastern Asia,0.8 to <0.2 Ma; Europe andsouthern Asia, 0.5? to <0.2 Ma)

1 Simple cores and “casual” flakes Early (Lower) Paleolithic Zhoukoudian (China, 1 to 0.25 Ma)Various (Europe, >1 to <0.4 Ma)Oldowan (Africa, 2.5 to <1.5 Ma)

∗Ka = thousand years ago; Ma = million years ago.

first fossils were found in Java in 1893 and termedPithecanthropus erectus. Most of the later findsin China and across Africa were given distinctivegeneric and specific names, but all are now usuallyconsidered local variants or subspecies of the singlespecies H. erectus. Some scholars still recognize ad-ditional species, especially H. ergaster for the earlierEast African populations, but here only one speciesis accepted. The major anatomical characteristics ofthis form are the following: a body of nearly modernform and proportions below the neck, topped by alow and slightly elongated skull with cranial capac-ity averaging 1100 ml (with a range of about 800–1300 ml), smaller teeth in a less projecting face thanAus. africanus or H. rudolfensis, large solid browridges, thick cranial bones, and no chin.

African populations. The earliest specimens are prob-ably from East Africa, dating to as much as 1.9 Ma.These populations might have continued to makesimple Oldowan artifacts, as no other form of stonetools is known at this time. There is evidence ofthe use of fire as early as 1.8 Ma and in scatteredsites thereafter. Acheulean bifaces (handaxes andcleavers) appear by 1.5 Ma, though only rarely indirect association with H. erectus fossils; evidencefor group hunting of large animals is rarer still andhas been questioned by some scholars. In addition,the eventual extinction of P. boisei in this time in-terval may have been caused by direct or indirectcompetition with the more advanced H. erectus.

The most important African H. erectus find of re-cent decades is a mostly complete skeleton of an ado-lescent male (its sex judged by pelvic bones as wellas by relative cranial robusticity), recovered on thewest side of Lake Turkana in 1984 (Fig. 8). Judgingby the pattern of dental eruption, he was probably9–10 years old at death, although an age of 11–12 hasalso been suggested. One still ambiguously answered

question is whether this species underwent the ado-lescent growth spurt which occurs in modern chil-dren of ages 12–14. Using formulas based on limbbone lengths of modern human adults and children,it was estimated that the “Turkana boy” might havereached an adult height of nearly 6 ft (183 cm) anda body weight of 150 lb (68 kg). Moreover, his bodyshape would have been slender or elongate, muchas seen in living people of the same region, as a re-sult of similar adaptation to a warm and dry climate.The development of brow ridges and other cranialstructures in the adolescent male and a larger skullfrom Olduvai contrasts markedly with other Turkanaskulls, indicating the persistence of strong sexual di-morphism.

A number of interesting new fossils thought to beH. erectus were found in East Africa in 1995–2002.A relatively complete cranium without a face fromLake Turkana and cranial fragments from anotherKenyan site represent quite small individuals. Thelatter was found in the same stratigraphic layer asmany bifacial Acheulean tools, good evidence thatthey were indeed the toolmakers. The first potentialH. erectus fossils from Ethiopia and Eritrea have alsobeen recently described, both from around 1 Ma.Measurements of the Ethiopian fossil indicate itssimilarity to both African and Asian H. erectus fos-sils. The Eritrean fossil, on the other hand, showsan increase in width of the back of the skull thatis also seen in modern humans. Together, thesenew fossils extend the known variation in AfricanH. erectus considerably.

Eurasian representatives. Homo erectus (perhaps asa result of increasing population size) spread fromAfrica into Eurasia through the Middle East, perhapsearlier than has previously been thought. New datesfor some long-known Javanese fossils suggestedages around 1.7 Ma, but these have been questioned.

474 Fossil humans

Fig. 8. Partial skeleton of adolescent male Homo erectusfrom West Turkana, Kenya, about 1.5 million years old.(Courtesy of and c© National Museums of Kenya)

Fig. 9. Crania and mandibles (in front view) of Homo erectus from Dmanisi, Georgia,reveal diversity in a single sample dated ca 1.7 Ma. (Courtesy of and c© Georgian NationalMuseum)

The most exciting new finds have come fromthe site of Dmanisi (Georgia), mainly from 1999 to2005 (Fig. 9). There are currently five crania, fourmandibles and a large number of postcranial ele-ments along with Mode 1 (Oldowan-like) tools andanimal bones dated to 1.8–1.6 Ma. Mode 1 and earlyMode 2 artifacts from Israel (as yet not definitivelyfound with human fossils) are slightly younger. Thus,H. erectus in the broad sense accepted here musthave been the first human species to leave Africa inlarge numbers, probably about 1.8 Ma, soon after itfirst appeared in the fossil record.

Fossils of this species may extend in mainland Asiato nearly 200,000 years ago, mostly associated withfauna from the warmer intervals in this time of alter-nating glacial climate. From central Java, Indonesia,come a series of cranial and dental remains spanningfrom perhaps 1.7 Ma to possibly 30 Ka. The earlierrange of time yielded several groups of skulls knownpopularly as Java man, including the first found mem-bers of this species; these specimens (and also thosefrom China) tend to be more derived or extreme intheir morphology than most African specimens, in-cluding the common presence of a thickened “keel”of bone down the midline of the braincase and amore projecting browridge. A dozen younger par-tial skulls from Ngandong, Java were long thoughtto be quite recent in age and once were placed inH. sapiens, but later study indicated close similarityto earlier Indonesian H. erectus, as they are now clas-sified. Their age remains uncertain, with some evi-dence suggesting a range between 400 and 100 Ka,while a dating attempt in 1996 arrived at the sur-prisingly young age of 50–30 Ka, which has beenquestioned on several grounds. If this is indeed acorrect date for these fossils, this would be the firstdemonstration of the coexistence of H. erectus withanatomically modern humans, known elsewhere inIndonesia and Australia at that date. Homo erectusspecimens continue to be recovered from Java, andone partial cranium was even recognized in 1999 ata New York City shop selling natural history speci-mens (Fig. 10). It was determined that it had beenfound in Indonesia some years earlier and illegally re-moved from that country; the shop’s owner returnedit to Indonesian authorities.

A fascinating find was made in 2003 on theIndonesian island of Flores in a cave called LiangBua, which has an archaeological record preservedfrom 200 Ka to the present. The fossils found there,including a well-preserved skull, additional lowerjaws, and postcranial bones, have been dated toabout 18 Ka. Dubbed “Hobbits” in the popular press,the remains are those of a small-brained (380 ml)and small-bodied (3 ft 3 in. or 1 m) hominin witha skull that resembles that of a small H. erectus.This led the discovers to conclude that a populationof H. erectus somehow reached the island overwater and became “dwarfed” after isolation, aphenomenon observed in other large mammalson islands; they named the fossils H. floresiensis.Other scientists have suggested that the remainsbelonged to pathological modern humans, which

Fossil humans 475

Fig. 10. Partial cranium of ?female Homo erectus fromSambungmacan, Java, Indonesia, age uncertain (probablybetween 500,000–100,000 years ago). This specimen wasidentified in a natural history store in New York City, yearsafter it was found by local workers in Indonesia. (Courtesyof and c© E. Delson; photo by C. Tarka)

seems unlikely given the number of individuals ofdiffering ages recovered. Further finds published inlate 2005 highlighted the long arms and other anoma-lous postcranial features of this species, which aredifficult to explain.

The first Chinese H. erectus, once called Pekingman, was found at Zhoukoudian, near Beijing, wherethey occupied a large cave during most of the periodbetween 500 and 250 Ka. Additional fossils of thisform are known from Yuanmou in the southwest,Lantian and Yunxian in the center, and Hexian andTongshan (Nanjing) in eastern China; some may be asold as 1–1.2 Ma. Archeological remains found withthese fossils are mainly of the simple Mode 1 vari-ety (no stone tools are unambiguously known fromlayers with human fossils in Java). However, morecomplex bifacial tools have been found in the BoseBasin of southern China, from about 0.8 Ma. Thesetools are comparable in overall form and complex-ity to Acheulian (Mode 2) tools found in Africa atthis time and predate Mode 2 tools in Europe. Thiscontradicts the long held hypothesis that only sim-ple Mode 1 tools were made in East Asia while morecomplex Acheulean tools were restricted to Africaand Europe at this time. Although there have beenclaims, no definite H. erectus fossils are yet knownfrom Europe, nor are archeological remains or moremodern humans unambiguously documented thereolder than about 800,000 years. The Dmanisi speci-mens from Georgia are the most ancient human rep-resentatives in the region, but the lack of furtherremains over the succeeding million years may indi-cate that Homo required further physical or culturaladaptations in order to survive in Europe proper.

Middle Pleistocene Homo. It has been suggestedthat the increased rigor of the glacial climate inEurope at this time was the impetus leading to theevolution of humans who seem to be physically moremodern in several ways than Afro-Asian H. erectus.These people were often termed early or archaicH. sapiens, or sometimes placed in their ownspecies, H. heidelbergensis. This view was accept-

able so long as the most ancient African represen-tatives of this group were poorly dated or younger,but studies in the later 1990s suggested some mod-ification. The earliest human fossils in Europe werelong thought to date to about 500 Ka in England(a tibia from Boxgrove) and Germany (the mandiblefrom Mauer near Heidelberg found in 1908). Thelong-known human fossils from Tighenif (previouslycalled Ternifine, in Algeria), dated to 800–700 Ka,were transferred out of H. erectus by some workersbecause they present at least one derived feature ofthe lower jaw. Moreover, the partial skull from Bodo(Ethiopia) was also dated older than 600 Ka, and sim-ilar South African fossils were estimated to be of com-parable age. These dates suggested that the earliestrepresentatives of these more derived humans mayhave lived in the northern half of Africa and perhapsevolved there from local H. erectus populations earlyin the Middle Pleistocene (780–125 Ka).

Then, in 1995, fragmentary human fossils and asso-ciated stone tools from the older levels at Atapuerca(Spain) were dated to about 800 Ka, implying thatarchaic H. sapiens appeared at about this date allover the western Old World. These specimens werenamed H. antecessor in 1997, and it was suggestedthat they represented the common ancestor of alllater human varieties. Meanwhile, in 1996, a cra-nium lacking the face was described from Ceprano,Italy, in a context suggestive of a date of about700 Ka or more. However, this date is not par-ticularly secure. A revised reconstruction claimsstrong similarity to H. erectus, otherwise unknown inEurope. In late 2005, stone tools (of both Mode 1and Mode 2 types) and animal fossils were describedfrom Pakefield in eastern Britain, a site estimated todate to 700 Ka; no human fossils are yet known, butthey are eagerly awaited.

All of these finds, combined with theoretical argu-ments about the best way to recognize and delimitspecies in the fossil record, have led to competinginterpretations of the number of species of Homoknown in the past million years. Some workers con-tinue to place all post-erectus fossils in “archaicHomo sapiens,” sometimes recognizing a variety oftemporal and geographic subspecies (such as theNeanderthals and anatomically modern humans).A few have gone so far as to include H. erectuswithin an over-enlarged H. sapiens. At the other ex-treme, some researchers accept between five andseven species in the same time period: H. anteces-sor, H. cepranensis (for Ceprano) H. heidelbergensis(either restricted to Europe or extended to Africaand even East Asia), H. rhodesiensis (for African Mid-dle Pleistocene populations), H. neanderthalensis,H. sapiens (restricted to anatomically modern hu-mans), and perhaps others. A possible middle groundwould be to (1) include the earliest of these Africanand European populations in one named group [forexample, antecessor, or perhaps mauritanicus (thename originally given to the Tighenif fossils), as sug-gested by J.-J. Hublin]; (2) combine all post-500 Kanonmodern European fossils in neanderthalen-sis (including heidelbergensis here); (3) group

476 Fossil humans

nonmodern African fossils younger than Tighenif inrhodesiensis; and (4) restrict sapiens to anatomicallymodern humans worldwide. These “groups” couldbe given specific or subspecific status, depending onthe theoretical model used. Recent work discussedbelow supports full species status for Neanderthalsbased on several lines of evidence, and all four groupsare here interpreted as full species. While both fos-sil and genetic evidence supports an African originfor modern humans, the evolutionary relationshipsamong all of these groups remains unclear.

Homo rhodensis and early representatives of H.neanderthalensis occur in Africa and Europe, re-spectively, between 650 and 250 Ka, thus contempo-raneous with H. erectus populations in eastern Asia.They share somewhat larger brains (for body size),smaller teeth, more expanded facial sinuses and oc-ciput (rear of the skull), but less robustness than inH. erectus. All these features are found in more ex-treme form in modern humans and in the late “clas-sic” Neanderthals (see below). In most areas, thesepeople still used Acheulean tools, but perhaps withgreater efficiency. It is likely that these “intermediate-grade” humans spread gradually eastward across theOld World, replacing late-surviving populations ofH. erectus everywhere by 200 Ka. These geographicvariants were not only distinct from H. erectus butalso from each other to a greater degree than istrue among living varieties or “races” of anatomicallymodern humans. In southern Africa, one craniumwas found at Broken Hill, now Kabwe, Zambia (for-merly Northern Rhodesia, hence the name Rhode-sian man), and broadly similar specimens are knownin South Africa (Saldanha and Florisbad), Tanzania(Ndutu), Ethiopia (Bodo), and Morocco (Salé andThomas quarries). These people made Acheulean orequivalent Mode 2 tools and apparently hunted biggame across most of Africa. In early 2006, a newpartial cranium probably belonging to this speciesfrom Gawis (Ethiopia, near the Gona Pliocene toolsite) was reported by Ethiopian and U.S. researchers

Fig. 11. Crania in frontal view of male Homo neanderthalensis from La Ferrassie, France(on right) and Amud, Israel, ca 50,000–60,000 years old. (Courtesy of and c© IsraelAntiquities Authority (Jerusalem), Musee de l’Homme (Paris) and E. Delson; photo byC. Tarka)

led by Sileshi Semaw. Its age is as yet uncertain butprobably in the 500–250 Ka interval.

Rare specimens from China appear to be younger,mainly dating to about 250–150 Ka. These includea nearly complete cranium from Dali, in centralChina, and a partial skull and skeleton from Jin-niushan (or Yingkou), in the northeast, as well asscattered, less complete remains. All of these fos-sils, especially Dali, are broadly similar to the Africanspecimens just mentioned, as well as to some ofthe earliest European H. neanderthalensis. OtherChinese specimens, such as Maba (from the south-east), and the central Indian Hathnora (or Narmada)fossil, are partial crania which are both younger (per-haps about 150–75 Ka) and more derived morpho-logically, although not in the direction of either earlyanatomically modern people or the contemporane-ous Neanderthals.

Neanderthals. The best known of the archaicvarieties are the Neanderthals, from Europe andwestern Asia. It now seems likely that this groupevolved locally in Europe from early Middle Pleis-tocene Homo via intermediate populations previ-ously called Homo heidelbergensis such as thoseknown from England (Swanscombe, ?Boxgrove),Spain (Atapuerca), France (Arago, Montmaurin), Ger-many (Mauer, Steinheim), and Greece (Petralona).They became adapted to the cold climates ofglaciated Europe, with prototypical Neanderthalanatomy well established by about 200 Ka. Duringthe warm interval about 120 Ka, they may havespread into the Near East and central Asia. In thecold glacial phase between 110 and 35 Ka, “clas-sic” (or extreme), cold-adapted Neanderthals wereabundant in cold northern parts of western andcentral Europe, while less extreme forms (perhapsmore like their immediate predecessors) inhabitedareas to the south and east. They were essentiallystocky humans, but had long, low skulls with aprojecting occipital region, large faces, teeth, andbrow ridges; and brains averaging 1500 ml in vol-ume (Fig. 11). Their limbs and trunks were heavilymuscled, indicating great strength, but many boneswere broken and healed during life. They madeMousterian tools (a variant of Middle Paleolithic orMode 3 flake-based tool kits), often lived in caves orwooden shelters where they controlled fire, huntedbig game, and had primitive religious beliefs, includ-ing burial of the dead with grave goods. There isintense argument among paleoanthropologists as tohow modern the Neanderthals were behaviorally,in terms of their stoneworking and hunting tech-niques and modes of foraging, whether plannedor merely ad hoc. For example, the frequent bro-ken bones may have resulted from hunting largegame at close proximity, rather than using projectiletechnology from a distance. Such controversies feedback into the question of whether the Neanderthalsare a distinct species or a distinctive subspecies ofH. sapiens. A related question is whether the Nean-derthals were in any way ancestral to anatomicallymodern humans, especially of Europe.

Recognition of a separate Neanderthal species im-plies an almost absolute reproductive isolation and

Fossil humans 477

lack of genealogical continuity, while the oppositeis true for most interpretations of Neanderthals asmembers of H. sapiens. The interpretation acceptedhere is that H.neanderthalensis is considered tohave been geographically and culturally isolated fromearly anatomically modern humans and their ances-tors, two independent but closely related lineages(species) evolving in parallel until they finally met,after which the former group soon became extinct.Indeed, current research suggests that differences inthe skulls between Neanderthals and modern hu-mans appeared early in individual growth, mean-ing that Neanderthal children looked quite differentfrom modern human children. Neanderthals also dif-fer in developing more rapidly than modern humans.In other words, they achieved their adult anatomyfaster than modern humans do. In addition, work byK. Harvati and colleagues published in 2004 showsthat the differences in skull anatomy between Nean-derthals fossils and modern humans are at the level ofseparate species when compared to other primates.Furthermore, when scientists extracted mitochon-drial DNA from the bones of several Neanderthalfossils they found that Neanderthals have someunique sequences of DNA when compared to mod-ern humans. Neanderthals, possibly due to local cli-matic differences, were probably isolated from otherhominins and as a result became quite distinct intheir development as well as their skull shape com-pared to modern humans. See NEANDERTHALS.

Spread of modern humans. One of the major foci ofrecent paleoanthropological research is the clarifica-tion of the area of origin and the early history of mod-ern humans, H. sapiens. The skull of this form is char-acterized by a small, nonprotruding face; small teethand brow ridges; a chin; and a high, rounded brain-case. There are no specimens of this type known(or even hinted at) anywhere in the world earlierthan about 200–150 Ka. New fossils representingearly modern humans were recovered from Herto,Ethiopia, in 1997 (published 2003) and dated 164–150 Ka. Aside from having larger faces than recenthumans, these are quite modern looking. Moreover,in 2005, the long-known fossils from Kibish (OmoValley, Ethiopia) were dated close to 200 Ka. Thisconfirms previous suggestions that about 200–100 Ka some fossils from eastern and southernAfrica (for example, one from Kibish and one froma late horizon at Laetoli) suggest the persistence ofa “Rhodesian-like” morphology, while others (for ex-ample, another from Kibish and several from DjebelIrhoud, Morocco) appear to be nearly modern. Twosomewhat younger sites in South Africa have pro-duced important additional modern human fossils.At Border Cave, a partial cranium and other frag-ments may date to nearly 90 Ka; they are clearlymodern in form, but their date is questionable. TheKlasies River Mouth caves, on the southern coast,have yielded a sequence of layers with good datesand archeological context; the human remains datedabout 100 Ka are scrappy but appear modern, witha chin, small brow ridges, and overall gracility. Suchgracilization, especially of the face, is a major fea-ture of the evolution of (and within) Homo sapiens.

In combination, these remains and other, less com-plete fossils indicate that early moderns were livingin sub-Saharan Africa by 150 Ka. Archeological re-mains dated 100–75 Ka in South Africa and D. R.Congo (ex-Zaire) indicate that at least some of thesepeople were making Mode 3 (MSA) tool kits withelements (such as bone harpoons, engraved bone,and minerals and shell beadwork) that do not ap-pear in Europe until after 30 Ka. See EARLY MODERN

HUMANS; PALEOLITHIC; PREHISTORIC TECHNOLOGY.From such a possible sub-Saharan origin, anatomi-

cally modern H. sapiens appear to have spread acrossthe Old World, differentiating into local races by 80–50 Ka. This view of human dispersal has receivedsupport from studies of the distribution pattern ofhuman mitochondrial DNA haplotypes (variants) andother genetic evidence. The majority of these stud-ies suggest that the major dichotomy in modernhuman population genetics is between Eurasians andAfricans. Such results fit well with the fossil evidencefor African versus Eurasian divergence about 100 Ka.Moreover, dates on early anatomically modern re-mains from Israel (Djebel Qafzeh and Skhul) docu-mented the presence of the ancestors of Eurasiansoutside Africa by about 110–90 Ka. This is especiallyintriguing because most Israeli Neanderthals havebeen dated to about 65–45 Ka, which is significantlyyounger than the early moderns and implies that theymay have overlapped in this region of the world.Even more complexity is implied by the strong sim-ilarity of Mousterian tool kits associated with suchdiverse human forms as the Neanderthals of Europeand Israel (and farther east), the early moderns fromIsrael, and the “premoderns” from Jebel Irhoud (Mo-rocco), but the implications of this cultural similarityare as yet unclear.

Found in the late 1970s, the youngest known Ne-anderthal skull comes from southern France andis associated with tools of the Châtelperronian in-dustry, a Mode 3 or 4 variety previously thoughtto have been made by H. sapiens. This specimendated to 34 Ka has been alternatively interpretedas the maker of these tools (possibly after contactwith Late Paleolithic moderns) or as evidence fordirect Neanderthal ancestry of moderns. In somecases, more fossils do not answer questions butcreate new ones. Sites in southern Spain and Por-tugal have yielded less complete Neanderthal fos-sils and Mousterian tools dated about 30 Ka, afterwhich modern H. sapiens was the sole form ofhuman to be found anywhere. One reason for thesuccess of H. sapiens may have been their greatertool-making efficiency, as documented by the Late(or Upper) Paleolithic Mode 4 blade-and-burin in-dustries. These people included large quantities ofworked bone in their tool kits (using chisel-likeburins to carve and engrave the bone), constructeddwellings of wood or of already fossilized animalbone, hunted large game, fished with harpoons,and in general behaved much like their living descen-dants. In many parts of the world, they also wore per-sonal ornamentation (jewelry) and engaged in artis-tic pursuits, including carving small animal statuesand perhaps calendars, as well as painting on the

478 Fossil primates

walls of rock overhangs and deep caves. See PALE-

OLITHIC; PREHISTORIC TECHNOLOGY.In 1999, Portuguese and American paleoanthro-

pologists described the remains of a 4-year-old child,buried near Lagar Velho, Portugal, about 25 Ka. Theburial pattern and most of the child’s morphology in-dicated links to the Gravettian culture, made by earlyanatomically modern people elsewhere in Europe atthis time. However, features of the lower leg bone(tibia) and lower jaw (the cranium was crushed andlater reconstructed) suggested similarity to Nean-derthals. The describers hypothesized originally andin a major publication in 2005 that this individualmight have been the result of hybridization betweenNeanderthals and moderns, but other researchersargued that the morphology was not that differentfrom what could be expected in a robust anatom-ically modern child and that hybridization wouldhave resulted in features intermediate between Ne-anderthals and moderns, not clear features of each.

Many names have been given to early modern hu-mans, especially in Europe, but these indicate onlyminor differences. The term Cro-Magnon derivesfrom several skeletons found in 1868 in Les Eyzies,France. They gave their name to a “race” said to occureither just in France or across most of Europe. Infact, Cro-Magnon people were already essentially Eu-ropeans, while early Africans are known from sitesin eastern and southern Africa. Australasia was colo-nized over water after about 70 Ka, with importantfinds at Keilor and Lake Mungo. New World Indianslikely originated from Siberia, by means of crossing aland bridge over what is now the Bering Strait. Manyhuman fossil remains are known in the Americas asfar back as 12 Ka, but some dates as old as 35 Kahave been obtained on archeological sites, indicat-ing that perhaps several crossings of the land bridgeoccurred.

In contrast to the above “Out of Africa” viewof human dispersal (based on the idea that mod-ern humans evolved in sub-Saharan Africa morethan 100,000 years ago from populations of archaicHomo) accepted here, a minority view (the “Multi-regional” hypothesis) interprets the fossil record todocument the nearly parallel origin of modern hu-mans in different regions of the Old World from a H.erectus ancestry. Each regional variety is said to pres-ent morphological characteristics linking archaic tomodern populations, while gene flow (migration andinterbreeding) between regions kept the geographi-cal varieties united in a single species at any one time.While most researchers reject a significant contribu-tion from Neanderthals or H. erectus to the modernhuman gene pool, there may still have been someinterbreeding among the species. See EARLY MOD-

ERN HUMANS; PALEOLITHIC; PREHISTORIC TECHNOL-

OGY. Eric Delson; Karen L. BaabBibliography. E. Delson (ed.), Ancestors: The Hard

Evidence, 1985; E. Delson et al. (eds.), Encyclope-dia of Human Evolution and Prehistory, 2d ed.,2000; J. G. Fleagle, Primate Evolution and Adapta-tion, 2d ed., 1998; W. C. Hartwig, The Primate Fos-sil Record, 2002; R. G. Klein, The Human Career:

Human Biological and Cultural Origins, 2d ed.,1999; J. H. Schwartz and I. Tattersall, The HumanFossil Record, vols. 1–4, 2001–2005; A. Walker andR. Leakey (eds.), The Nariokotome Homo erectusSkeleton, 1993; P. F. Whitehead, W. K. Sacco and S. B.Hochgraf, A Photographic Atlas for Physical Anthro-pology, 2005.

Fossil primatesExtinct members of the order of mammals to whichhumans belong. All current classifications divide theliving primates into two major groups (suborders),but zoologists differ as to whether the tarsier (Tar-sius) should be classified with the lower primates(lemurs, lorises, and bushbabies) or the higher pri-mates (New and Old World monkeys, greater andlesser apes, and humans).

All primates have a common origin that, however,is not reflected in the universal possession of a suiteof diagnostic features. The order as a whole has beencharacterized in terms of showing a group of progres-sive evolutionary trends, notably toward predomi-nance of the visual sense, reduction of the senseof smell and associated structures, improved grasp-ing and manipulative capacities, and enlargement ofthe higher centers of the brain. Among the extantprimates, the lower primates more closely resem-ble forms that evolved relatively early in the historyof the order, while the higher primates represent agroup more recently evolved (Fig. 1).

A classification of the primates, as accepted here,is shown on the next page.

Early primates. The earliest primates are placed intheir own semiorder, Plesiadapiformes (as contrastedwith the semiorder Euprimates for all living forms),because they have no direct evolutionary links with,and bear few adaptive resemblances to, any group ofliving primates. However, the chewing teeth and thelocomotor anatomy of these fossil forms sufficientlyresemble those of later primates to demonstrate thecommon origin of the two groups. Best known fromthe Paleocene Epoch, around 65–55 million yearsago (Ma), and found in both the Old World and theNew World, most plesiadapiforms retained clawedhands and feet. However, in 2002 it was shown that atleast one species, Carpolestes simpsoni, possessedan opposable great toe with a nail, as in euprimates(Fig. 2). Plesiadapiforms also had rather small brainscompared to their body size, possessed large spe-cialized front teeth, were probably arboreal in habit,and may have been adapted for terminal branch feed-ing, as recently suggested. These animals are knownfrom fossil deposits as far north as Ellesmere Island,in Arctic Canada, which, during the Paleocene, wascovered by the subtropical forest stretching contin-uously from western North America across a land-locked North Atlantic into western Europe.

Eocene primates. There is no known plesiadapi-form that is a satisfactory candidate for the ancestryof the fossil primates of modern aspect typical of thesucceeding epoch, the Eocene (55–34 Ma). Often