Current events

Mzalendo Kibunjia Pliocene and Pleistocene archeological Department of Anthropologv, Rutgers sites west-o+ Lake Turkana, Kenya l~niuersity, New Brunswick, New Jersey 08903, U.S.A. and National Museums of Ke~va, P.O. Box 40658, ./Vairobi, K-ecva

Htlkne Roche Nztional Museums of Kenya, P.O. Box 40658, Nairobi, Kenya and Centre ;Vational de la Recherche Scient$que, ERA 28 du CRA, 92190 Meudon, France

Francis H. Brown Department of Geology and Geophysics, Uniwrsity of Utah, Salt Lake CiQ, LTtah 841 l-7, U.S.A.

Richard E. Leakey I;enya Wildlife Seruices, P. 0. Box 40241, .Vairobi Kenya D _ Journal ofHuman Evolution (1992) 23,431-438

Geographical setting

The Nachukui Formation lies between present day Lake Turkana and the Murua Rith and Labur Ranges to the west. Sedimentary deposits in the area consist of patchy Miocene strata and fossiliferous Pliocene and Pleistocene deposits. Woody shrubs, acacia trees and doum palms presently exist along the ephemeral streams and the broad interfluvial plains are covered with short grasses and xerophytic forbs. The region is now dissected by west-east draining seasonal streams and this configuration may not have changed since antiquity, except that during Pliocene and Pleistocene times streams probably drained into a larger river system or into a different lake than today (Brown & Feibel, 1988).

The formation was deposited in the Turkana Basin, an integrated paleogeographic unit in which all Pliocene and older Pleistocene deposits are intimately related (Feibel et al., 199 1) . These strata make up the Omo Group, which includes the Shungura Formation in the lower Omo Valley and the Koobi Fora Formation east of the lake. The Nachukui Formation has an aggregate thickness of 730 m and is subdivided into eight members using widespread volcanic tuffs as marker beds (Harris et al., 1988a,6). These are (with the approximate age ranges given parenthetically), the Lonyumun (>4 Ma), Kataboi (4-3.35 Ma), Lomekwi (3.35-2.5 Ma), Lokalalei (2.5-2.35 Ma), Kalochoro (2.35-1.9 Ma), Kaitio (1.9-1.65 Ma), Natoo (1.655 1.3 Ma) and Nariokotome ( 1.3%< 0.7 Ma) Members.

Archeological sites

The archeological sites of Lokalalei (GaJh 5), Kokiselei (FxJh 5) and Naiyena Engol (FxJh 6) are named for the ephemeral streams along which they occur. Stratigraphically, the

0047-2484/92/ 11043 1 + 08 $08.00/O 0 1992 Academic Press Limited

432 M. IUBUNJEA ET AL.

’ / LOWARENGAK

_

Figure I. Map of the West Turkana area where the four archeological sites are located. Many other archaeoic@d occurrences presently under invesrigation were fiund during surveys in the upper four members ofthe Nachukui Formation.

Lokalalei site lies near the base of.the Kalochoro Member and the two others lie within the Kaitio Member. The site of Nadung’a (FxJh 71, which is younger than Lokaialei and

Kokiselei, lies within the Nariokotome Member (Figure 1). About 35 m of section are exposed along the northern branch of the Lokalalei channel

where Lokalalei site is situated (Figure 1). The strata dip gently to the northwest and include a volcanic ash layer about 5 m above the base of the local section, which has been named the Ekalalei Tuff. The Ekalalei Tuff has been confidently correlated with Tuff F-l of the Shungura Formation (Harris et al., 19883) and lies 9 m beneath the archeological site. Several other tuffs exposed in the vicinity (e.g., the Lokalafei TuR and the Kokiselei Tuff> equivalent to Tu& D and E ofthe Shungura Formation) give confidence to the stratigraphic placement. Regional mapping demonstrates that the site lies far below the next dated tuff in

ARCHEOLOGICAL SITES, WEST TURKANA 433

the section, the KBS Tuff (1.8810.02 Ma; McDougall, 1985). The section between the Lokalalei site and the KBS Tuff includes a fine-grained sequence of lacustrine sediments that correlates with submembers G-14 to G-26 of the Shungura Formation. The beginning of this lacustrine episode of deposition is believed to have begun at about 2.0 Ma. As deposition appears to have been continuous above the Ekalalei Tuff at the Lokalalei site, it is likely that it has an age only slightly younger than that of Tuff F (2.36 f. 0.04 Ma; Feibel et al., 1989). Immediately underlying the Ekalalei Tuff at the site is a conglomerate with volcanic clasts that represent deposition by a small stream entering the basin from the Murua Rith range to the west. The artefacts are enclosed in silty claystones at the top of an upward-fining cycle deposited by a large meandering river. Paleogeographically, the site appears to have lain near the interface between ephemeral streams draining the western hills and the main course of the ancestral Omo River.

Kokiselei is a large seasonally intermittent channel and the Kokiselei site is situated in rugged hills on the south side of its major northern branch. The strata dip very gently to the west and include an altered volcanic ash layer near the base, identified as the KBS Tuff of the Koobi Fora Formation ( =Tuff H-2 of the Shungura Formation) by analysis of the glass phase (Harris et al., 19883). This layer can be traced discontinuously northward through the Naiyena Engol and Kalochoro drainage systems. Since the Kokiselei site overlies the KBS Tuff, it is younger than 1.88 Ma. No tephra layers have been found above this site in the immediate vicinity. However, a section between the KBS Tuff and the Nariokotome Tuff has been pieced together in similar (and correlative) deposits exposed between Kalochoro and Nariokotome, where the Nariokotome Tuff has been dated at 1.33 kO.05 Ma (Brown et al., 1985; Feibel et al., 1989). Assuming a constant rate of sediment accumulation between the two dated levels, an age of 1.84 Ma is suggested, but until the stratigraphy is investigated further we will use a defensible age estimate of 1.8-tO.l Ma. The strata near the site are dominated by conglomeratic sandstones with pebble and cobbles of volcanic rocks and mollusk-packed sandstones, which are interpreted as high-energy shoreline deposits. The archeological site lies in silty claystones that were probably deposited in a low area between the lake shore to the east and an alluvial plain extending westwards towards the mountains.

The Naiyena Engol site is located on the ridge which divides the Naiyena Engol drainage from that of Kalochoro. It lies 17.4 m above the KBS Tuff in a stratum and setting similar to that of Kokiselei and would appear to be only slightly younger, again falling in the range of 1.8 &- 0.1 Ma. The principal difference between the two sites is that the site of Naiyena Engol lies in a thicker sequence of fine-grained strata than that of Kokiselei.

The Nadung’a site lies about 25 m above the Nariokotome Tuff (1.33 Ma) and about 25 m below another tuff unknown elsewhere in the basin. Pumices excavated at the site during the test excavation have not yet been dated, but some are compositionally similar to pumices from the Silbo Tuff (0.74 Ma). The site is thus clearly younger than 1.3 Ma and possibly less than 0.74 Ma in age. The Nadung’a site lies in the same sort ofsilty claystones as do the sites at Kokiselei and Naiyena Engol.

Only small test excavations were made to confirm that the artefacts are in situ in each of the sites, but each site is clearly more extensive. The excavated surfaces in the four sites are not equivalent: 14 m* at Lokalalei, 5 m* at Kokiselei, only 2 m2 at Naiyena Engol; and 3 m* at Nadung’a. Artefacts are quite scarce at Lokalalei and vertically scattered through 28 cm within the sediment, whereas they are more abundant and dispersed only through 25 and 18 cm, respectively, at Kokiselei and Naiyena Engol. In spite of this dispersal, it is likely that there is only a single archaeological level in each of these three sites. There are probably two

434 M. KIBUN,JIA ETAL.

Table 1 Fauna1 remains from the Lokalalei, Kokiselei and Naiyena Engol sites

Lokalalei Kokiselei Naiyena

Engol

12=45 n=39 n=9 Pisces + + + Reptilia

Crocodylia + + Serpentes + Chelonia + +

Mammalia Primates

Cercopithecidae indet. + Colobinae indet. +

Perissodactyla Equidae indct. +

Artiodactyla Suidae indet. +

Kolpochoerus aff. limnetes + Hippopotamidae indet. + + Giraffidae indet. + Bovidae Bovidae indet. (small size) + + Bovidae indet. (medium size) Bovidae indet. (large size) + Alcelaphini indet. + +

Dnmaliscus sp. + Antilopini indet. + Neotragini indet. + Hippotragini indet. + Tragelaphini + +

levels at the site ofNadung’a which is rich in artefacts. Ofthese, the lower level is undisturbed and the artefacts are dispersed only through 15 cm. The integrity of the archeological sites in any case depends on the depositional conditions (stream bank or floodplain at Lokalalei, lakeshore at Kokiselei, Naiyena Engol and Nadung’a) and on post-depositional processes. During excavation, termite nests were noticed at Lokalalei, and small cracks affect the fine sediment at Kokiselei. Nearly half of the artefacts at Kokiselei lie vertically or sub-vertically in the sediment, while 87% of those at Nad-ung’a and 70% of those at Lokalalei and Naiyena Engol are horizontal or subhorizontal. The lithic material is very fresh at Lokalalei, but some of the material from Kokiselei and Nadung’a, and almost all of the material from Naiyena Engol, shows surface alteration. Cobbles and pebbles were locally available near all sites in small stream channels draining the western hills and similar material likely occurred on beaches not far from the Kokiselei and Naiyena Engol sites.

Bone remains are very fragmented in the three older sites and only a small percentage of the fragments could be identified taxonomically (Table 1). Preservation of bone differs considerably between the three sites. At Lokalalei and Naiyena Engol the bones are dense and gray to bluish black, whereas those at Kokiselei are white and chalky. Bone from the surface at Kokiselei shows percussion marks, trampling marks and carnivore marks, but only one excavated specimen from this site exhibits possible cut marks. No fossil bones were excavated at Nadung’a.

ARCHEOLOGICAL SITES, WEST TURKANA 435

Table 2 Lithic assemblages from the four sites

Lokalalei Kokiselei Naiyena

Engol Nadung’a

Excavated surface cores Whole flakes Broken flakes and fragments Very small flakes and chips (< 1 cm) Retouched flakes Worked or broken pebbles Unworked pehbles Total

14m2 5m? 7 5

14 76 4 41 4 191

_ _

_ 2 2 1

31 316

2m? 3m2 3 6

54 175 18 64 27 165 _

3 4

109

8 3

422

The lithic assemblages

Additional excavation is necessary before a detailed interpretation of the sites can be pro- posed. However, it is probable that our small excavations at the Kokiselei and Nadung’a sites correspond to stone knapping localities. This is indicated by an assemblage ofonly a few cores but hundreds ofvery small flakes and chips (Table 2). At Nadung’a, a few bifaces were found on the surface around the site, but none were found in situ. This site is much younger than the three other sites, so comparison with them is probably not appropriate. We do note, however, that in Nadung’a the percentage ofbroken flakes is quite low. Comparison ofspecific features of the lithic material at the three older sites-Lokalalei, Kokiselei and Naiyena Engol--is especially worthwhile because it may provide insight about the technological competence of the hominids who used the material.

The lithic assemblages consist mainly of cores and whole or broken flakes and fragments, but vary from the oldest site to the younger ones (Table 2). At Lokalalei, there are only a few flakes (Figure 2), perhaps because of the conditions of deposition, but a good sample of very crude cores (Figure 3) made from lava cobbles of medium diameter (m= 120 mm) that are particularly thick (up to 115 mm). Similar cobbles are present in the beds ofmodern streams in the area and a possible bedrock source for these is exposed about 5 km to the west in the mountains along Kokiselei. The number of scars on the cores varies from very few (5-6) to more than 20. Two cores are flaked unifacially, one with only one direction ofpercussion and the second with several directions ofpercussion. The other cores are thick cobbles, truncated by several series of flakes removed by strong, randomly applied, blows (e.g., Figure 3). The purpose of the flaking appears to have been the removal of flakes and not the shaping of the block per se: the number of scars from flake removal varies and no real cutting edge exists on the core. Careful observation shows that many flaking accidents occurred (hinged flakes, plunging flakes, broken flakes). Therefore, most of the flakes produced were either small and poorly formed, or broken. Most of the raw material is fine-grained lava that is quite suitable for flaking, so the poor flakes did not arise because the material was inappropriate. Such rocks, though, are brittle and breakable when not properly worked, so the poor flakes may result from expedient tool manufacture and may also indicate ill-organized practice and poor skill (Callahan, 1979; Johnson, 1979). By contrast, the few cores collected from the younger sites, Kokiselei and Naiyena Engol, are more like the classical chopper-cores, with a cutting edge (either deliberate or not) obtained by a sequence ofcontiguous flake removal. This may

436 M. KIBUNJIA ETAL.

0 I TE 3 40

Figure 2. Flakes from Lokalalei Site showing typical accidental breaks that occurred while flaking. From top to bottom: LA87 Q48/1, LA 87 R48/1, LA87 R47/1.

indicate a more sophisticated conceptual schema and greater manual skill, even though the number of broken flakes is still noteworthy.

Conclusions

Our observations at these sites in West Turkana are consistent with those made on other archaeological sites of Pliocene and Pleistocene age, despite differences between the raw materials utilized for tool-making. Lokalalei can be compared with the sites at Kada Gona in the Hadar region (Roche & Tiercelin, 1980; Harris, 1983) and with sites in the Shungura Formation (Chavaillon, 1976; Merrick & Merrick, 1976), which are older than 2 Ma. Kokiselei and Naiyena Engol are closer in age to the Oldowan sites in Bed I and Lower Bed II at Olduvai (Leakey, 1976) and to early sites in the Koobi Fora Formation (Isaac & Harris, 1976). We have noted that the tools from Lokalalei (2.35 Ma) appear to be less refined than those at Kokiselei and Naiyena Engol (1.8 Ma). If th is observation is correct, then it would seem that the tool-making abilities of hominids changed markedly sometime during this

ARCHEOLOGICAL SITES, WEST TURKANA 437

Figure 3. Core no. LA87 S46/ 1 from the Lokalalei Site, made from a thick fine-grained lava cobble. Almost all the scars from flake removal show a flaking accident.

half-million year interval (cf. Roche, 1989). This coincides with other behavioral changes in early hominids as well, for example, their subsistence behavior, their movements within a territory and the areas they appear to have occupied for extended periods (Harris & Capaldo, in press). Considering that Homo habilis (Leakey et al., 1964; Howell, 1978) is known in the Turkana Basin only after 2 Ma (Harris et al., 19886)) our archaeological data again seem to confirm that stone tools were made before H. habilis emerged. The recent report of a fossil of the genus Homo from Baring0 (Hill et al., 1992) that may be as old as the site of

Lokalalei will no doubt give rise to speculation that the tools from Lokalalei were made by the taxon. There are, however, still no fossils attributed to Homo of this age from the Nachukui Formation. Whether or not the increased technological proficiency apparent in the tools of the Kokiselei and Naiyena Engol sites can be clearly correlated with the emergence of H. hub&s requires further investigation, but it is a possibility.

Acknowledgements

We thank the Government of Kenya and the Governors of the National Museums of Kenya. This research was funded by the National Geographic Society, the National Museums of Kenya, the Leakey Foundation and Mr Robert Brownlee. We also thank Bw. Kamoya Kimeu and his team for invaluable help in the field, John Kimengich and Jean Philip Brugal for fauna1 determinations and Mary Salvaggio for bone mark identifications. Comments on the manuscript by Curtis Marean, Nicholas Toth, Catherine Perles and other colleagues are gratefully acknowledged.

438 M. KIBUNJIA ET AL.

References

Brown, F. H., McDougall, I., Davies, T. & Maier, R. (1985). An integrated Pli~Pleistocene chronology for the Turkana Basin. In (E. Delson, Ed.) Ancestors: The HardEvidence, pp. 82-90. New York: Alan R. Liss.

Brown, F. H. & Feibel, C. S. (1988). “Robust” hominids and Plio-Pleistocene paleogeography of the Turkana Basin, Kenya and Ethiopia. In (F. E. Grine, Ed.) Evolutionary Histq ofthe “Robust” Australopithecines, pp. 325-341. New York: Aldine de Gruyter.

Callahan, E. (1979). The basics ofbiface knapping in the Eastern Fluted Point Tradition: a manual forflintknappers and lithic analysts. Archaeol. E. jY. Am. 6, I-213.

Chavaillon, J. (1976). Evidence for the technical practices of early Pleistocene hominids Shungura Formation, Lower Omo Valley, Ethiopia. In (Y. Coppens, F. C. Howell, G. Ll. Isaac & R. E. F. Leakey, Eds) Earliest Man and Enuironments in the Lake Rudolf Basin, pp. 565-573. Chicago: University of Chicago Press.

Feibel, C. S., Brown, F. H. & McDougall, I. (1989). Stratigraphic context of hominids from the Omo Group deposits: northern Turkana basin, Kenya and Ethiopia. Am. 3. p/y. Anthrop. 78,595-622.

Feibel, C. S., Harris, J. M. & Brown, F. H. (1991). Neogene paleoenvironments of the Turkana Basin. In 0, M. Harris, Ed.) Koobi Fora Research Project, Volume 3, Stratigraphy, Artiodact_yls and Paleoenuironments, pp. 321-346. Oxford: Clarendon Press.

Harris, J. W. K. (1983). Cultural beginnings: Pli~Pleistocene archaeological occurrences from the Afar, Ethiopia. Afr. Archaeol. Reu. 1,3-32.

Harris, J. M., Brown, F. H., Leakey, M. G., Walker, A. C. & Leakey, R. E. (1988a). Plio-Pleistocene hominid bearing sites from west of Lake Turkana, Kenya. Science 239,27-33.

Harris, J. M., Brown, F. H. & Leakey, M. G. (19886). Stratigraphy and paleontology of Pliocene and Pleistocene localities West ofLake Turkana, Kenya. Nat. His. Mu. L.A. County. Co&. Sci. 399, l-128.

Harris, J. W. K. & Capaldo S. D. (in press). The earliest stone tools: their implications towards an understanding of the activities and behaviours of late Pliocene hominids. In L’utilisation de Poutil chez lesprimates (Colloque de la Fondation Fyssen, Versailles 1988).

Hill, A., Ward, S., Deino, A., Curtis, G. & Drake, R. (1992). Earliest Homo. Nature 355,7 19-722. Howell, F. C. (1978) Hominidae. In (V. J. Maglio & H. B. S. Cooke, Eds) Evolution ofAfrican Mammals, pp. 154-248.

Cambridge: Harvard University Press. Isaac, G. Ll. & Harris, J. W. K. (1976). Archaeology. In (R. E. Leakey & M. G. Leakey, Eds) Koobi Fom Research

Project 1, Oxford: Clarendon Press. Johnson, J. K. (1979). Archaic biface manufacture: production failures, a chronicle ofthe misbegotten. Lithic Tech. 8,

25-35. Leakey, L. S. B., Tobias, P. V. & Napier, J. R. (1964). A new species ofthe genus Homo from Olduvai Gorge. Nature

202,7-g. Leakey, M. D. (1976). Olduvai Gorge, III Cambridge: Cambridge University Press. McDougall, I. (1985). K-Ar and ““Ar/jgAr dating of the hominid-bearing Plio-Pleistocene sequence at Koobi Fora,

Lake Turkana, northern Kenya. Geol. Sot. Am, Bull. 96, 159-175. Merrick, H. V. & Merrick J. P. S. (1976). Archeological occurrences ofearlier Pleistocene age from the Shungura

Formation. In (Y. Coppens, F. C. Howell, G. Ll. Isaac & R. E. F. Leakey, Eds) Earliest Man andEnvironments in the Lake RudolfBasin, pp. 574-584. Chicago: University of Chicago Press.

Roche, H. (1989) Technological evolution in the early hominids. OS&4 14,97-98. Roche, H. & Tiercelin, J. J. (1980). Industries lithiques de la formation Plio-Pleistocene d’Hadar: campagne 1976.

In (R. E. Leakey & B. A. Ogot, Eds) Proc., VIIIth Panafrican Congr. of Prehistory and Quaternary studies (19771, pp. 194-199. Nairobi: The International Louis Leakey Memorial Institute ofAfrican Prehistory.