When Does Homo Appear in the Omo Turkana Basin

An answer to this question involves several interrelated and very complex issues. One of the major problems is that Homo itself remains a poorly defined genus, and recognizing this taxon in the fossil record is thus highly problematic (Wood and Collard, 1999; Wood, 2009). The earliest specimens attributed to Homo on the basis of dental characters occur in Member E of the Shungura Formation, about 2.4 Ma (Suwa et al., 1996), and the lower Kalochoro Member of the Nachukui Formation, about 2.3 Ma (Prat et al., 2005). One of the problems with this first appearance datum (FAD) is that most of these specimens are isolated teeth, and therefore provide little information on most of the traits frequently used to define the genus (e.g., large endocranial volume). Another type of uncertainty relates to the fact that hominins were rare elements of Plio-Pleistocene mammalian faunas (Bobe et al., 2002). The Shungura fossil record collected by the American contingent of the International Omo Research expedition includes 22,335 specimens, of which 147 are identified as hominin. These numbers would indicate that hominins made up about 0.6% of the paleofauna (Fig. 15.2), but numerous collection and taphonomic factors can bias the number of collected specimens in relation to the original living faunas. A major portion of the Shungura record was collected by Gerald Eck using a well-defined and systematic methodology (Bobe and Eck, 2001; Eck, 2007). For example, he set out to collect all mammalian astragali along the Shungura deposits: a total of 601 astragali were collected, including 10 primates, but not a single astragalus was identified as hominin. This fact alone indicates that hominins were rare elements of the fauna near the environments of deposition of the paleo-Omo River. Although rare, hominins were clearly present in the paleo-Omo landscape. Mammalian mandibles were also systematically collected by Eck's team: 773 belong to identifiable mammalian families, and only 5 of these are hominin. Systematically collected mandibles thus indicate that hominins made up about 0.65% of the specimens on the surface of the Shungura Formation deposits (Fig. 15.2). All five ofthese hominin mandibles are identified as Paranthropus; no mandible attributed to Homo was found in these systematic surveys. In contrast, there are 147 cercopithecid mandibles (19% of the mammalian sample), and 96 of these (12.4% of the mammalian sample) can be identified to the genus Theropithecus, a primate that is taphonomically comparable to hominins (in terms of body size and morphology). These numbers do not necessarily indicate the actual abundance of hominins (or primates) in the Plio-Pleistocene of the lower Omo Valley, but they do indicate that hominins were rare and that Homo was particularly rare. Assessments of hominin paleobiology, as well as estimates of first and last appearances need to take this fact into consideration. Large samples are needed to detect the presence of rare taxa in a given landscape. When it comes to rare species such as those of early Homo, absence of evidence is not evidence of absence.

We can estimate the uncertainty associated with a first appearance datum (FAD) by considering the abundance of a

M

G

a

c

"te

"es

CJ

c

CJ

,—

c cececeJS

"ce^ttcL OOOOOOhÍI

ce ce ce re

ce ce ce ce

"cscd'cs'cs'cs'csE-^'cs jaijajajajatafc cni^cncncncnOcn ta

oo oo O

CD CD cD CD

ta ta o cj

LOOOOOC^l oooo oooo

ta

es .n

Ji

i

ta

St=

=

=

=

=

E—

E—

E—

E—

J3

J3

"-'

ín

J3

J3

-C

OJ:

C\J

CQ

-C

c/o

e

C--

CQ

re

1—

O:

í^ c

C

CJ

CJ

CJ

CJ

C5

c^

^

^

c

^

E—

p

P

p

P

P

E-

en ooooo en ooooo ooooo i—ILOO r— cor— ojejej cj ooc^

jz ce

o CE

h lE

C ta

en cc

es C

LD

r-

c^

ex

S:

S:

S:

S:

X

C

<5

<5

CJ

CJ

p

E-

E^

p

p

p

p

Table 15.1 (continued)

Omo

West Turkana

East Turkana

GPTS Age (Ma)

Error

Dated unit (Fm)

Method

Tuff D

Lokalalei

Lokalalei Tuff

Lokalalei Tuff

2.52

0.05

Tuff D (Sh)

K-Ar. Ar-Ar

Tuff C9

Emekwi Tuff

2.581

Emekwi (Nk)

GPTS Gauss-Matuyama

Unit C9

Lomekwi

Basal Burgi Mb

2.581

Unit C9 (Sh)

GPTS Gauss-Matuyama

Burgi Tuff

2.64

0.05

Burgi Tuff (KF)

K-Ar

Tuff C4

Ingumwai Tuff

2.74

0.08

Tuff C4 (Sh)

Stratigraphie scaling

Tuff C (alpha)

Hasuma Tuff

2.85

0.08

Tuff C (Sh)

Stratigraphie scaling

Tuff BIO

2.95 3.04

0.05

GPTS Kaena top

Ninikaa Tuff

3.08 3.11

0.03

Ninikaa (KF)

K-Ar, Ar-Ar GPTS Kaena base

U14

Karo

3.2

U14 (Us)

Allia Tuff

3.22

0.05

Allia Tuff (KF)

Stratigraphie scaling

Unit B2 (U12)

Tulu Bor

3.22

Unit B2 (Sh)

GPTS Mammoth top

Unit B1

3.33

Unit B1 (Sh)

GPTS Mammoth base

Toroto Tuff

3.32

0.02

Toroto (KF)

K-Ar. Ar-Ar

TuffB-beta

Tulu Bor Tuff (beta) Burrowed bed

Tulu Bor Tuff

3.40 3.53

0.03

Interpolation

Tuff A (U6)

Lokochot Tuff

Lokochot

Lokochot Tuff Wargolo Tuff

3.594 3.77

Tuff A (Sh) VT-3 (Maka)

GPTS Gilbert-Gauss

Usno 1

Kataboi

MoitiTuff

Moiti

MoitiTuff

3.94

0.04

Moiti Tuff (Nk)

Ar-Ar

Topernawi Tuff

Topernawi Tuff

3.96

0.03

Topernawi (Nk)

Ar-Ar

Kataboi Basalt

4.05

0.06

Basalt (Nk)

K-Ar

Kanapoi Tuff

4.07

0.02

KT (Kp)

Ar-Ar

Usno Basalt

4.10

0.06

Basalt (Us)

K-Ar. Ar-Ar

Mursi Basalt

Lonyumun

4.2

0.2

Basalt (Mu)

GPTS = Geomagnetic Polarity Time Scale Nk Nachukui Formation Kp Kanapoi Formation Nw Nawata Formation KF Koobi Fora Formation Sh Shungura Formation Us Usno Formation Mu Mursi Formation

Ages given of submembers reflect the age at the base of the unit, unless noted otherwise. Colors are used to separate adjacent members within each formation.

Table 15.2 Number of specimens among bovid tribes in geological members of the Nachukui, Koobi Fora, Usno, and Shungura Formations (Nachukui data from Harris et al., 1988)

Formation

Member or unit

Bovid tribes

Total

%Alcelaphini +Antilopini

Time (midpoint)

Aepycerotini

Alcelaphini

Antilopini

Bovini

Hippotragini

Reduncini

Tragelaphini

Nachukui

Kataboi

4

2

2

0

0

3

3

14

28.6

3.7

Nachukui

Lomekwi

55

58

12

8

0

68

15

216

32.4

3

Nachukui

Lokalalei

3

5

1

4

0

2

1

16

37.5

2.5

Nachukui

Kalochoro

6

15

12

4

0

36

7

80

33.8

2.15

Nachukui

Kaitio

11

20

1

6

0

11

26

75

28.0

1.75

Nachukui

Natoo

4

10

10

7

0

13

4

48

41.7

1.5

Nachukui

Nariokotome

3

4

0

2

0

6

2

17

23.5

1.3

Koobi Fora

Lokochot

16

10

4

4

0

10

25

69

20.3

3.5

Koobi Fora

Tulu Bor

4

14

7

25

0

96

39

185

11.4

3.1

Koobi Fora

Upper Burgi - KBS

114

231

128

78

19

512

186

1.268

28.3

1.9

Koobi Fora

Okote

6

35

7

15

0

71

33

167

25.1

1.5

Usno

B(L)

128

4

3

28

0

4

50

217

3.2

3.3

Shungura

BP

49

7

1

24

0

63

32

176

4.5

2.9

Shungura

C

94

5

3

73

3

74

178

430

1.9

2.7

Shungura

D

56

6

1

15

0

30

66

174

4.0

2.5

Shungura

E

91

9

0

13

1

56

117

287

3.1

2.4

Shungura

F

141

25

5

15

0

69

87

342

8.8

2.36

Shungura

G(L)

461

51

4

38

2

684

375

1.615

3.4

2.2

Shungura

G(U)

24

13

9

2

0

29

2

79

27.8

2

Shungura

H

7

12

3

2

0

125

4

153

9.8

1.8

Shungura

J

6

8

0

1

0

39

0

54

14.8

1.65

Shungura

K

5

15

2

1

0

27

1

51

33.3

1.5

Shungura

L

4

11

1

2

0

42

2

62

19.4

1.35

Total

1.292

570

216

367

25

2.070

1.255

5.795

13.6

Fig. 15.2 Abundance (number of specimens) of mammalian mandibles systematically collected from the Shungura Formation deposits. This distribution provides an estimate of the relative abundance of mammalian families in the Shungura deposits. There are five hominid mandibles in a sample of 773 mammalian specimens. Thus, hominids make up about 0.65% of the sample.

Fig. 15.3 The earliest occurrence of a species in the fossil record provides an upper limit to the timing of its true origination or migration into the area being sampled. The first appearance datum (FAD) should always be evaluated in relation to earlier samples. If the fossil record prior to the species' FAD is abundant and continuous, then the FAD has a high likelihood of representing a true origination or migration event. If the fossil record prior to the species' FAD is poor and discontinuous, then the FAD may be an artifact of sampling.

Fig. 15.3 The earliest occurrence of a species in the fossil record provides an upper limit to the timing of its true origination or migration into the area being sampled. The first appearance datum (FAD) should always be evaluated in relation to earlier samples. If the fossil record prior to the species' FAD is abundant and continuous, then the FAD has a high likelihood of representing a true origination or migration event. If the fossil record prior to the species' FAD is poor and discontinuous, then the FAD may be an artifact of sampling.

taxon once it appears in the fossil record, and the abundance and distribution of samples prior to that FAD. An observed FAD in the fossil record provides an upper limit to the species time of origination, but the true origination may be significantly earlier than the observed event (Fig. 15.3).

A first appearance preceded by a very sparse fossil record could simply indicate that the taxon has not been found in the sparse samples, even though it may have been present in the landscape. A first appearance of an abundant taxon preceded by a series of large samples without evidence of that taxon would provide a high level of confidence that the FAD truly reflects an origination or migration event, not an artifact of sampling. It should be noted that a FAD could indicate either in situ evolution or migration into the area being sampled. The only way to distinguish between these two phenomena would be to obtain samples from large areas, regional or continental in scale, that include the likely sources of the species in question. The current distribution of the fossil record in Africa, with a disproportionate contribution from the East African Rift Valley, precludes firm conclusions regarding immigration or speciation events. Nevertheless, we can use methods designed to evaluate the uncertainty of FADs depending on the distribution of samples within a region and the abundance of the species of interest. Here we use the methods developed by Koch and colleagues to place 95% confidence intervals on the FADs of Homo (Barry et al., 2002; Koch, 1987; Koch and Morgan, 1988). We estimate the probability (Pi) that Homo originated or migrated to the area being sampled by

Fig. 15.2 Abundance (number of specimens) of mammalian mandibles systematically collected from the Shungura Formation deposits. This distribution provides an estimate of the relative abundance of mammalian families in the Shungura deposits. There are five hominid mandibles in a sample of 773 mammalian specimens. Thus, hominids make up about 0.65% of the sample.

where n is the abundance of Homo over its range (number of specimens), m is the total number of fossil mammals in the collection over the same range, and r is the number of specimens in successive intervals prior to the first occurrence (FAD) of the genus. Thus, the ratio n/m provides a measure of relative abundance. Following Barry et al. (2002), we use a P. value of at least 0.8 to determine the 95% confidence interval. The same method can be applied to last occurrence data (LADs), but the focus here is on origination rather than extinction.

In the sample from the Shungura Formation, Homo appears in Member E, at about 2.4 Ma, represented by specimen L. 26-1, a right lower M1 (Suwa et al., 1996). Member E is preceded by relatively large samples of fossil mammals in Member D (with 946 specimens dated from 2.52 to 2.4 Ma), Member C (with 3301 specimens dated from 2.85 to 2.52 Ma), and Member B (with 1997 specimens derived mostly from the uppermost units dated to about 2.95 Ma). These earlier Shungura samples in the time range from 2.4 to nearly 3.0 Ma, and totaling 6,244 fossil mammals have not provided evidence of Homo. Thus, the earliest record of Homo at 2.4 Ma in the Shungura Formation may be very close in time to the true origination or migration event (in Fig. 15.4a the 95% confidence interval is narrow). But could we recognize Homo on the basis of isolated teeth prior to 2.4 Ma? Defining and recognizing Homo on the basis of an incomplete fossil record remains a daunting task. Nevertheless, we know that the Homo and Paranthropus lineages diverged sometime in the Pliocene. The earliest specimen assigned to Paranthropus is a right lower M2 (specimen L. 62-17) from Shungura Member C-5 (Suwa et al., 1996). Tuff C4 is dated to 2.74 Ma (Feibel et al., 1989) and the Paranthropus specimen derives from the sedimentary unit above it. Thus, Homo and Paranthropus clearly had separate evolutionary trajectories by 2.7 Ma, and possibly considerably earlier. If this is the case, the earliest record of Homo in

Fig. 15.4 First appearance datum (FAD indicated by arrow) of Homo in the (a) Shungura, (b) Koobi Fora, and (c) Nachukui Formations. Horizontal brackets denote the 95% confidence interval associated with the Homo FADs. See text for details.

the Shungura Formation at about 2.4 Ma would represent the migration of the genus into the lower Omo Valley.

At Koobi Fora there are exceptional specimens of unambiguous early Homo: KNM-ER 1470, KNM-ER 1813 and KNM-ER 3733 among the best known (Wood, 1991). The earliest of these are from the Upper Burgi Member of the Koobi Fora Formation, and date to slightly less than 2 Ma (Feibel et al., 1989; Spoor et al., 2007). They are roughly coincident with the earliest stone artifacts at East Turkana (Fig. 15.4b). But at Koobi Fora, we encounter the problem that there is a major unconformity spanning a few hundred thousand years below the Upper Burgi Member. Sediments from the Lower Burgi Member have yielded few fossils, and in the upper part of the Tulu Bor Member there is a sample of only 204 fossil mammals. Thus, prior to the FAD of Homo at Koobi Fora there is a long time interval, almost 1 Myr, with a relatively sparse fossil record. Thus, at Koobi Fora, the FAD of Homo at near 2 Ma has a large margin of error because of geological circumstances and a sparse fossil record prior to that first appearance (this is therefore depicted by a wide confidence interval in Fig. 15.4b).

The west side of Lake Turkana has produced spectacular hominins such as KNM-WT 15000, KNM-WT 17000 and KNM-WT 40000 (Brown et al., 1985; Leakey et al., 2001; Walker et al., 1986), but the earliest published specimen attributed to Homo is an isolated right lower M1 (KNM-WT 42718) from the base of the Kalochoro Member and dated to about 2.3 Ma (Prat et al., 2005). Compared to the Shungura Formation, the number of fossils from the Nachukui Formation is relatively small and therefore the earliest appearance of Homo at near 2.3 Ma in West Turkana has a wider 95% confidence interval (Fig. 15.4c). However, it is noteworthy that the archeological record at West Turkana is very close in age to the earliest Homo specimens, as is the case in East Turkana and the lower Omo Valley (Fig. 15.4). The approach taken here could be modified by archeologists to address confidence intervals associated with the earliest occurrences of lithic technologies, but the focus of this paper remains on the fossil bones.

The data discussed here indicate that the critical time period for understanding the origin and dispersal of Homo is between about 3 and 2.4 Ma. By 2.4 Ma there is evidence of Homo in the Omo-Turkana Basin and elsewhere in East Africa (Hill et al., 1992; Kimbel et al., 1997; Suwa et al., 1996). But profound changes in hominid adaptation seem to occur with the emergence of Homo erectus (sensu lato) after 2 Ma (Bramble and Lieberman, 2004; McHenry and Coifing, 2000; Wood and Collard, 1999; Wood, 2009). What was the environmental context in which these evolutionary processes took place? The mammalian fauna from the Omo-Turkana Basin provides a rich source of information on the environmental and ecological conditions that existed during the time when Homo first appeared and the subsequent emergence and dispersal of Homo erectus.

What Does the Mammalian Fossil Record Tell Us About the Environmental and Ecological Conditions Associated with the Emergence of Homo?

Multiple lines of evidence indicate that East African Plio-Pleistocene environments were complex and dynamic, having been composed of varying proportions of forest, woodland, bush, and grassland (Bobe, 2006; Bonnefille, 1995; Kingston, 2007; Kingston et al., 1994; Reed, 1997; Wynn, 2004). On broad scales, Pliocene African climate fluctuated between wet and dry conditions with an apparent overall trend toward greater aridity (deMenocal, 1995, 2004). The record of paleosol carbon isotopes shows a shift toward C4 grasslands during the latest Pliocene and Early Pleistocene (Levin et al., 2004; Wynn, 2004). This Plio-Pleistocene shift is also indicated by data from biomarkers recovered from site 231 in the Gulf of Aden (Feakins et al., 2005), although the two records are not entirely congruent (Fig. 15.5). While the paleosol record shows the shift to C4 vegetation at around the Pliocene/Pleistocene boundary, the biomarker record from the Gulf of Aden shows an earlier shift between about 3.4 and 3.0 Ma (Feakins et al., 2005). The explanation of this discrepancy is likely due to the larger area sampled by the wind-blown biomarker data. These residues of terrestrial

Fig. 15.5 (a) Plant biomarker data from the Gulf of Aden (site 231) depicting an increase in C4 vegetation at about 3.2 Ma followed by a further shift after 2 Ma (from Feakins et al., 2005). (b) East African paleosol carbonate data showing an increase in C4 biomass beginning about 2 Ma (data from Wynn, 2004). (c) Relative abundance of mammals indicative of seasonally arid grasslands in the lower Omo Valley showing a moderate increase in grassland-adapted mammals at about 2.5-2.4 Ma and more pronounced peaks after about 2 Ma (Data from Bobe and Behrensmeyer, 2004).

plants were recovered from marine records in the Gulf of Aden, and their source may have spanned an area including much of northeastern Africa as well as the Arabian Peninsula, while the paleosol data clearly derived from localized terrestrial sequences at hominin-bearing sites. An important conclusion to be drawn from these carbon isotope studies is that eastern African Pliocene/Pleistocene vegetation was complex and included an increasing component of C4 grasslands, but this expansion of grassland habitats likely occurred in different parts of East Africa at different times.

Analyses of fossil mammals present a pattern of change that supplements the carbon isotope record. The record from the Shungura Formation, for example, shows remarkable faunal stability during much of the Late Pliocene, but taxa indicative of open grasslands become slightly more abundant at about 2.5 Ma and more significantly so after 2 Ma (Bobe and Behrensmeyer, 2004) (Fig. 15.5). The Omo-Turkana data also show that different parts of the basin had different habitats. Thus, bovids indicative of seasonally arid grasslands were consistently more abundant in the East and West Turkana areas (Koobi Fora and Nachukui Formations) than in the Omo area (Shungura Formation) (Bobe et al., 2007). The diversity of habitats in the Pliocene-Pleistocene Turkana Basin is demonstrated here by comparing the relative abundance of Alcelaphini and Antilopini across different areas (Fig. 15.6). These antelopes are associated with grassland and bushland, seasonally arid habitats (Greenacre and Vrba, 1984). Omo, West Turkana, and East Turkana show very different percentages of alcelaphines and antilopines during the late Pliocene, but display some convergence during the latest

Fig. 15.6 Relative abundance of Alcelaphini + Antilopini (as a percentage of all Bovidae) in the three main regions of the Turkana Basin. High values suggest open, seasonally arid environments dominated by grasslands. The Omo appears to have been consistently more closed than other parts of the Turkana Basin, but all three regions show increases in grasslands, although with fluctuations, after 2 Ma.

Fig. 15.6 Relative abundance of Alcelaphini + Antilopini (as a percentage of all Bovidae) in the three main regions of the Turkana Basin. High values suggest open, seasonally arid environments dominated by grasslands. The Omo appears to have been consistently more closed than other parts of the Turkana Basin, but all three regions show increases in grasslands, although with fluctuations, after 2 Ma.

Pliocene and early Pleistocene, beginning at about 2 Ma. This would indicate grasslands and bushland were becoming more prominent throughout the basin beginning at about 2 Ma, although woodlands and forest fringing the paleo-Omo River would have remained an important part of the vegeta-tional landscape. Previous analyses controlling for tapho-nomic factors, depositional environments, and collection biases in the Omo region indicate that these faunal changes represent true biological phenomena (Bobe and Eck, 2001).

Among suids, species of Nyanzachoerus and Notochoerus decline precipitously toward the Pliocene/Pleistocene boundary, when species of Metridiochoerus and Kolpochoerus become numerically dominant (Fig. 15.7). Although the Notochoerus lineage shows increasing adaptation to a diet of grass over time (Harris and Cerling, 2002), it may have succumbed to competition from a large array of grazers that became widespread during the earliest Pleistocene. Among suids, Metri-diochoerus hopwoodi, M. compactus, and M. modestus are all highly hypsodont, and have their first appearances near the Pliocene/Pleistocene boundary. Likewise, species of Kol-pochoerus, which are characterized by significant elongation of the third molars, have their first appearances during the early Pleistocene (e.g., K. majus,K. phacochoeroides) or become dominant elements of the early Pleistocene fauna (e.g., K. hesel-oni) (Cooke, 1997; Harris and Cerling, 2002; White, 1995).

Other mammalian indicators of grasslands show a similar pattern of increasing dominance during the latest Pliocene or earliest Pleistocene. The genus Equus, for example, a quintessential grazer of the African savanna, first appears in the Omo record at about 2.3 Ma (at the base of Member G), where it comprises about 30% of all specimens of the family Equidae (other species belong to the genus Eurygna-thohippus). After about 1.8 Ma, Equus makes up about half or more of all equid specimens (Fig. 15.8).

Among cercopithecids, Theropithecus brumpti is the dominant monkey in the Omo from about 3 to 2.4 Ma (Members B through D). It is also numerically dominant at Koobi Fora during Tulu Bor times (after 3.4 Ma), along with large bodied colobines such as Rhinocolobus turkanaensis, Paracolobus mutiwa, Cercopithecoides kimeui, and C. williamsi (Jablonski and Leakey, 2008). In Shungura Member E, at 2.4 Ma, T. oswaldi begins to replace T. brumpti, which becomes extinct by 2 Ma (Eck et al., 1987). The extinction of T. brumpti is followed by the decline of large bodied colobines in the Koobi Fora Formation by 1.5 Ma, but during this time T. oswaldi continues to thrive (Leakey et al., 2008).

These results point to consistent trends in the fauna over time. They do not necessarily provide environmental r econstructions of particular time periods, but they do document complex patterns of ecological change that likely influenced the behavior of hominin populations during the Plio-Pleistocene.

Was this article helpful?

0 0

Post a comment