- Email: [email protected]
Confirmation of a late middle Pleistocene age for the Omo Kibish 1 cranium by direct uranium-series dating
Journal of Human Evolution 63 (2012) 704e710
Contents lists available at SciVerse ScienceDirect
Journal of Human Evolution journal homepage: www.elsevier.com/locate/jhevol
Confirmation of a late middle Pleistocene age for the Omo Kibish 1 cranium by direct uranium-series dating Maxime Aubert a, *,1, Alistair W.G. Pike b, Chris Stringer c, Antonis Bartsiokas d, Les Kinsley a, Stephen Eggins a, Michael Day d, Rainer Grün a a
Research School of Earth Sciences, The Australian National University, Canberra ACT 0200, Australia Department of Archaeology and Anthropology, University of Bristol, 43 Woodland Road, Bristol BS8 1UU, UK Department of Palaeontology, The Natural History Museum, London SW7 5BD, UK d Department of History and Ethnology, Democritus University of Thrace, P.O. Box 217, 69100 Komotini, Greece b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 January 2012 Accepted 14 July 2012 Available online 5 September 2012
While it is generally accepted that modern humans evolved in Africa, the specific physical evidence for that origin remains disputed. The modern-looking Omo 1 skeleton, discovered in the Kibish region of Ethiopia in 1967, was controversially dated at w130 ka (thousands of years ago) by U-series dating on associated Mollusca, and it was not until 2005 that AreAr dating on associated feldspar crystals in pumice clasts provided evidence for an even older age of w195 ka. However, questions continue to be raised about the age and stratigraphic position of this crucial fossil specimen. Here we present direct Useries determinations on the Omo 1 cranium. In spite of significant methodological complications, which are discussed in detail, the results indicate that the human remains do not belong to a later intrusive burial and are the earliest representative of anatomically modern humans. Given the more archaic morphology shown by the apparently contemporaneous Omo 2 calvaria, we suggest that direct U-series dating is applied to this fossil as well, to confirm its age in relation to Omo 1. Ó 2012 Elsevier Ltd. All rights reserved.
Keywords: Oldest anatomically modern humans Homo sapiens Laser ablation Africa Ethiopia
Introduction The three fossil humans found in 1967 in the Kibish Formation of the lower basin of the Omo River, southern Ethiopia, come from different locations and stratigraphic contexts (Butzer, 1969; Butzer et al., 1969; Leakey, 1969). Omo 1, a partial skull and associated skeleton was excavated at site KHS in Member I. Omo 2, a calvaria, was a surface find reportedly at site PHS, 2.5 km away on the other side of the Omo River, but allocated to Member I by correlation (however, see below). Omo 3, consisting only of cranial fragments, was apparently derived from Member III. Omo 1 and 3 represent anatomically modern Homo sapiens, but Omo 2 is more archaic in cranial morphology (Day, 1969; Day and Stringer, 1991). Radiocarbon dates for the upper part of the Kibish Formation suggested that Members IeIII lay beyond the practical limits of the method, and this was consistent with uranium-series age estimates of
* Corresponding author. E-mail address: [email protected] (M. Aubert). 1 Present address: Centre for Archaeological Science, School of Earth and Environmental Sciences, University of Wollongong, Wollongong 2522, Australia. 0047-2484/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jhevol.2012.07.006
130 5 ka (thousands of years ago) on Etheria shells from Member I, correlated by Butzer (1969) with the stratigraphic position of the Omo 1 skeleton. In turn this suggested that the Omo 1 skeleton was potentially the oldest known example of the modern form of Homo sapiens. However, serious questions were subsequently raised about the reliability of these U-series dates on mollusc material (see e.g., Smith, 1992 and following discussion). Renewed excavations at the original locality have produced further hominin fossils, including some from the Omo 1 skeleton, recovered in situ (Fleagle et al., 2008; Pearson et al., 2008a, b), and clarified the location and stratigraphic placement of the Omo 2 find (McDougall et al., 2005, 2008; Brown and Fuller, 2008; Feibel, 2008). Associated dating studies correlated AreAr determinations on feldspar crystals in pumice clasts from Member I with regional fluvial cycles and Mediterranean sapropels, suggesting a revised age of between 172 ka and w195 5 ka for Omo 1 and 2 (McDougall et al., 2005, 2008; Brown et al., 2012). This work was in turn re-evaluated by Millard (2008), who proposed an age somewhat younger than 190 ka, while Klein (2009) raised the possibility that the Omo 1 skeleton was in fact a much younger intrusive burial. Direct dating work on the Omo 1 skeleton is thus of critical importance in confirming or denying that this fossil may
M. Aubert et al. / Journal of Human Evolution 63 (2012) 704e710
represent the oldest known example of anatomically modern H. sapiens. We have applied U-series analysis to gain further insights to the age of the fossil. U-series dating of bones is seriously compromised by the fact that bones can accumulate large amounts of uranium following their deposition into sediments. A range of models have been developed to account for this uranium uptake and provide a basis for open system dating. The diffusion-adsorption (D-A) model was developed by Millard (1993) and Millard and Hedges (1996), and refined by Pike (2000) and Pike et al. (2002). It is based on laboratory experiments and assumes a continuous diffusion of uranium from the outside of a bone or tooth towards the interior, and that the partitioning between the bone and solution (groundwater) and the U concentration in the solution are constant. The bone is treated as a homogeneous medium. Under constant conditions, the cross sections of bones that conform to the D-A diffusion model are expected to have both u-shaped Uconcentration and apparent U-series age profiles, with the apparent ages at the surface being closest to the correct age of the sample. Deviations from such ideal profiles can be explained either by leaching or changes in the U concentration in the solution. The D-A model was recently refined by Sambridge et al. (2012). Their Diffusion-Adsorption-Decay (DAD) model expands the D-A model for diffusion of 234U and its decay during the diffusion process. For a given volume in a bone, the DAD model postulates that 234U is continually resupplied by diffusion. As a result, 234U/238U ratios change little over time, consequently DAD model ages are somewhat older than comparative D-A results. All age results presented here are based on the assumptions underlying the DAD model. Samples We have dated two small (w4 g) parietal fragments (A and B) of the Omo Kibish 1 skull that could not be fitted onto the cranial reconstruction and that were retained in London, first at St. Thomas’s Hospital Medical School, and then the Natural History Museum. Fragment A was initially analysed along a single laser track while fragment B was later analysed along a series of laser track on two different planes. For more details on the analysed surfaces see Results and Discussion below. Experimental Laser ablation elemental U, Th and U-series isotope analyses were carried out at the Australian National University. For the experimental set-up, see Eggins et al. (2003, 2005), and for applications of U-series dating on human fossils, see Grün et al. (2005, 2006, 2008). U and Th concentrations were derived from repeated measurements of the NBS-610 standard, U-isotope ratios from the dentine of a rhinoceros tooth from Hexian (sample 1118, see Grün et al., 1998). The ANU Neptune mass spectrometer has only one ion counter that is required for the measurement of both 230Th and 234U. As a consequence, each track is run twice, first for the estimation of the 230 Th/238U ratio, then for the 234U/238U ratio. When U-series data are derived from spot analysis, two closely adjacent holes are drilled. It turns out that the 234U/238U ratio usually changes little within a bone, so that this analytical approach does not introduce any significant additional errors when compared with laser ablation analyses using multiple ion counters (e.g., Hellstrom, 2006; Grün et al., 2010; Duval et al., 2011). To remove surface contamination, the tracks and spots were first cleaned with the laser using a wider spot size. None of the measurements were affected by detrital Th. 230Th/232Th activity ratios were well above 100, mostly above 1000.
705
Measurements were carried out in 1024 ms cycles. Along the tracks, an apparent U-series age was calculated for each cycle. These had 2-s errors in the 20% range. To reduce the noise of the data scatter, a sliding five point average was applied. The errors of the individual age estimates are not shown. Both samples had large pores, some filled with secondary minerals, the U-concentrations dropped considerably (often to less than 1 ppm) from the average of around 30e50 ppm. The low Uconcentrations lead to very large errors in the calculation of Useries isotope ratios, and consequently the mean ages scatter widely. These regions were excluded in the age profiles that lead to much more coherent results when using a five point sliding average. In previous experiments, we found that there are differences in the measured 230Th/234U ratios in the range of 3% depending on the direction of the laser tracks. The cause for this is at present not resolved. To alleviate any problems relating to the motion of the sample in the ablation cell during isotope measurement, we carried out the spot-analyses, which does not require any sample movement. The holes were drilled with the laser set to 137 mm in diameter for 100 s. Ages were calculated from the integrated data and have individual 2-s errors in the range of 5%e6%. The highest 230 Th/234U ratios of the track analyses are about 3% higher than the corresponding spot results. Age calculations During the alpha decay of 238U, a parent atom experiences a strong recoil. In minerals, this recoil leads to a weakened bond between the remaining atoms and the crystal lattice. When minerals are exposed to weathering, 234U atoms are preferably leached. This process leads to the well-known fact that surface and ground waters have excess 234U over 238U (Cherdyntsev, 1971). This has some major implications for the dating of bones. In closed systems, for example for speleothems or corals, U-series ages are calculated from the measured 230Th/234U and 234U/238U ratios. The decay of excess 234U is taken into account in the age calculation. Bones, however, are an open system for uranium and for a given volume, the 234U/238U ratio remains more or less constant by the continuing re-supply of 234U from the outside. As a result, most bones have rather uniform 234U/238U ratios. Indeed, all 234U/238U ratios of Omo fall into a narrow range of between 1.41 and 1.46. Assuming that the 234U/238U ratios remain constant over time, bone ages are calculated from the present 230Th/234U ratios. For more details see (Sambridge et al., 2012). Results and discussion Initially, sample A was analysed along a single laser track. The apparent U-series ages and U-distributions indicated a rapid Uuptake with a best age estimate of 98þ8 6 ka (Fig. 1). We had submitted a paper containing these results in 2005, but unfortunately this coincided with the publication of the Ar/Ar results of McDougall et al. (2005), who had obtained an age estimate on a tuff from Member 1, underlying the human remains, of 195.8 1.6 ka and an estimate of 103.7 0.9 ka for tuff in the upper part of Member 3, about 50 m above the former sample. Considering that the first result was obtained from the same sedimentary unit in which the human remains were found while the second was derived from a unit that was deposited at least two sedimentary cycles later, McDougall et al. (2005) made a convincing case that the age of Omo 1 should be close to 195 ka. Whilst our results did not contradict the findings of McDougall et al. (2005), they pointed to the basic problem of U-series dating of bones, namely that the results are usually minimum age estimates, and that it is difficult to
706
M. Aubert et al. / Journal of Human Evolution 63 (2012) 704e710
Figure 1. Age and U-concentration along a single laser track from sample A measured in 2005. The apparent U-series ages and U-distributions indicated a rapid U-uptake with a best age estimate of 98þ86 ka.
impossible to estimate by how much the U-series results underestimate the correct age of the sample (Pike et al., 2002). In the game of finding the oldest modern humans, a direct minimum age of 98 ka would have been sensational some 20 years ago (e.g., Grün and Stringer, 1991), but in light of a 195 ka Ar/Ar age on the sediments, the U-series results were dismissed out of hand. Detailed U-series mapping of a Neanderthal tooth from Payre (Grün et al., 2008) indicated that bones and teeth may contain domains with very different apparent U-series ages, the oldest being close to the expected age range. The sample B had a nearly flat surface, presenting a cross section through the bone. To minimise destruction, this surface was evened out with fine sand paper. Six tracks were run across this bone (Fig. 2). The U-concentration profiles along the six tracks are either flat or show some increased concentrations close to the outer surfaces. This is more or less expected from the DiffusionAdsorption model (Pike et al., 2002). However, the apparent Useries data are not entirely symmetrical, the apparent U-series ages at the top of the fragment are somewhat older than at the base. Furthermore, close to the surfaces, the apparent U-series results are often younger, particularly where the U-concentrations increase (e.g., Track 2, between cycles 240 and 260, or Track 5, between cycles 1 and 10). This points to a secondary overprint of uranium. Track 1 generally replicated the earlier results with ages around 100 ka. However, the apparent U-series age estimates gradually increase, particularly in the upper part to the right, where apparent age estimates of around 170 ka are reached along Tracks 5 (around cycle 60) and 6 (between cycles 100e120). This may either point to an earlier U-uptake phase, or to U-leaching. In the scanned surface, no evidence for leaching is apparent, which would be associated with lower U-concentrations. Nevertheless, uranium could have been lost through the exposed surface, i.e., perpendicular to the measurement plane. It was decided to cut the fragment about 2 mm below the original surface with a diamond wire saw. The cutting width was around 110 mm. The subsequent U-series analyses were carried out on the underside of the first plane (Fig. 3). For easier comparison, the image is flipped. The U-concentrations in this second plane are somewhat lower than in the upper plane. Ignoring the pores, the Uconcentration profiles are either flat or increase towards the outer surfaces. Similar to the upper plane, the apparent U-series age estimates indicate secondary overprints in some domains close to the surface (e.g., Tracks 1, 5 and 6, close to the ends of the scans in
Fig. 3). Otherwise, the apparent age profiles are either flat or Ushaped, with a tendency of older ages towards the upper part of the fragment (compare beginnings and ends of the tracks). At the upper right hand side (particularly Tracks 9 and 10), ages of more than 200 ka were obtained. There is no evidence for lateral U-leaching. The general agreement between the apparent U-series estimates in the upper and lower planes indicates that the old ages were not produced by U-leaching along a perpendicular direction. When running the ablation cell, it was observed that some differences in the 230Th/238U ratio may occur depending on the absolute track direction, i.e., when the same track is analysed after being rotated by 90 , the results may be somewhat different. This, to our knowledge, does not affect the relative results along a track, or between parallel tracks, as long as they are measured in the same direction. To address this problem, spot analyses were carried out in the domain that yielded the oldest apparent U-series ages (see dotted square in Fig. 3, and Fig. 4). As mentioned above, a U-series age calculation using spots consists of two separate analyses. Therefore, only about half of the holes can be used for a U-series calculation (see Fig. 4A). However, spot analyses have the advantage of not being dependent on movements within the cell, and are averaged over much longer measuring times, resulting in significantly smaller individual errors. The calculated mean ages are shown in Fig. 4B. The spots provided old ages on the same vicinity where the tracks also yielded the oldest apparent U-series age estimates. As mentioned above, this domain did not show any signs of U-leaching expressed in lower U-concentrations. In addition, the 234U/238U ratios change from around 1.44 to 1.46 close to the outside (upper part of larger square in Fig. 4A) to 1.41 to 1.42 further to the inside. At the same time, the apparent 230Th/234U ages decrease from around 160 to 180 ka close to the outside to 140 to 150 ka further inside. This is consistent with the predictions of the DAD model (Sambridge et al., 2012), i.e., a function of U-uptake. When Uleaching occurs, the 234U/238U ratios in the affected volumes generally decrease due to preferential leaching of 234U (e.g., Duval et al., 2011). The average ages are younger than those observed along the tracks, but are based on significantly larger volumes. A detailed 3D study of U concentrations and U-series isotope ratios in an early Pleistocene horse tooth (Duval et al., 2011) showed that small-scale uranium redistribution may lead to significant changes in the respective 230Th/238U ratios. This may well explain some of the high apparent ages in the inner square, or the low age of 92 ka at the base between tracks 10 and 11. Averaging over larger areas will address the problem of micro-redistribution. Concentrating at the end of tracks 9 to 11 (inner square in Fig. 4), an average age of 171 16 ka is obtained, averaging a much larger area (outer square in Fig. 4) results in 162 14 ka. The much larger errors of the averages compared to the errors of the individual results are an indication of U-mobilisation. Assuming that Omo1 has indeed an age of close to 195 ka (McDougall et al., 2005), these older domains in the bone approximate this age closely. Our detailed U-series analyses also point to the deficiency of Useries dating of bones: if the analysis is confined to a small section, e.g., a single track, severe underestimations may occur. These underestimations are probably due to changes in the U-concentrations in the percolating waters or mineralogical changes in the bones that allow the adsorption of higher U-concentrations with time. Furthermore, many bone sections may well not contain any domains that have preserved the original isotopic signature. As a result, most U-series age estimations on bones have to be regarded as minimum age estimates, even when D-A or DAD model predictions are apparently fulfilled (e.g., Fig. 1). On the other hand, data sets complying with the D-A or DAD models will not overestimate the age of a bone.
M. Aubert et al. / Journal of Human Evolution 63 (2012) 704e710
707
Figure 2. Age and U-concentration along individual tracks 1e6 from sample B (upper plane). The surface was obtained by slightly abrading the original surface with fine sand paper. Track 1 generally replicated the earlier results from sample A with ages around 100 ka. However, the apparent U-series age estimates gradually increase, particularly in the upper part to the right, where apparent age estimates of around 170 ka are reached along Tracks 5 (around cycle 60) and 6 (between cycles 100e120). The analysed surface area measures approximately 30 mm by 15 mm.
708
M. Aubert et al. / Journal of Human Evolution 63 (2012) 704e710
Figure 3. Age and U-concentration along individual tracks 1e12 from sample B (lower plane). The surface was obtained by cutting the fragment about 2 mm below the original surface with a diamond wire saw. The subsequent U-series analyses were carried out on the underside of the first plane. Similar to the upper plane, the apparent U-series age estimates gradually increase towards the upper part of the fragment. At the upper right hand side (particularly Tracks 9 and 10), ages of more than 200 ka were obtained. The general agreement between the apparent U-series estimates in the upper and lower planes indicates that the old ages were not produced by U-leaching along a perpendicular direction. The analysed surface area measures approximately 30 mm by 15 mm.
M. Aubert et al. / Journal of Human Evolution 63 (2012) 704e710
709
Acknowledgements Aspects of this study were supported by Australian Research Council Grants DP0664144 (Grün et al.) ‘Microanalysis of human fossils: new insights into age, diet and migration’ and DP0666084 (Roberts et al.) ‘Out of Africa and into Australia: robust chronologies for turning points in modern human evolution and dispersal’. Chris Stringer is a member of the Ancient Human Occupation of Britain project, funded by the Leverhulme Trust, and his research is supported by the Calleva Foundation. References
Figure 4. Results of drilling holes with the laser into the domain that yielded the oldest apparent U-series ages (see dotted square in Fig. 3). A: 234U/238U and 230Th/238U ratios. B: Age estimates. Note that the average ages are somewhat younger than those observed along the tracks, but are based on significantly larger volumes.
Implications The results presented here illustrate that a wide range of apparent U-series ages can be generated on a particular sample and that careful evaluation of the data is required. The younger apparent ages obtained here appear to be related to higher Uconcentrations pointing out to a later overprint of uranium. This is further illustrated by the higher U-concentrations and younger age estimates in the upper plane (sample B) compared with the lower plane. It also appears that the domain that yielded the oldest apparent U-series ages was not produced by U-leaching so that the minimum age estimate for this sample is w155e187 ka. In spite of the methodological complications involved in Useries analysis of bones, we expect that laser ablation U-series dating will lead to further critical insights into the timing of modern human evolution, and we recommend that this technique should be applied to the morphologically distinct Omo 2 calvaria to test the stratigraphic and chronological relationship between the Omo 1 and Omo 2 fossils. If their contemporaneity can be established, this will further demonstrate the great morphological variability displayed among early H. sapiens fossils in Africa, and suggestions of more complex scenarios for the evolutionary origins of our species in that continent (Stringer, 2007, 2012; Gunz et al., 2009). Notwithstanding those issues, our new direct dating of the Omo 1 cranium seem to confirm its middle Pleistocene age within Member I of the Kibish Formation, and refute any suggestion that it represents a much later intrusive burial. Presently, the Omo 1 specimen is the earliest known representative of the anatomically modern skeletal pattern.
Brown, F.H., Fuller, C.R., 2008. Stratigraphy and tephra of the Kibish Formation, southwestern Ethiopia. J. Hum. Evol. 55, 366e403. Brown, F.H., McDougall, I., Fleagle, J.G., 2012. Correlation of the KHS Tuff of the Kibish Formation to volcanic ash layers at other sites, and the age of early Homo sapiens (Omo I and Omo II). J. Hum. Evol. 63 (4), 577e585. Butzer, K.W., 1969. Geological interpretations of two Pleistocene hominid sites in the Lower Omo Basin. Nature 222, 1133e1135. Butzer, K.W., Brown, F.H., Thurber, D.L., 1969. Horizontal sediments of the lower Omo Valley: the Kibish Formation. Quaternaria 11, 15e29. Cherdyntsev, V.V., 1971. Uranium-234. Israel Program for Scientific Translations, Jerusalem. Day, M.H., 1969. Omo human skeletal remains. Nature 222, 1135e1138. Day, M.H., Stringer, C.B., 1991. Les restes crâniens d’Omo Kibish et leur classification à l’intérieur du genre Homo. L’Anthropologie 94, 573e594. Duval, M., Aubert, M., Hellstrom, J., Grün, R., 2011. High resolution, three dimensional mapping of U-concentrations and U-series isotopes in an Early Pleistocene equid tooth from Fuente Nueva-3 (Orce, Andalusia, Spain) using laser ablation ICP-MS. Quartern. Geochronol. 6, 458e467. Eggins, S.M., Grün, R., McCulloch, M.T., Pike, A.W.G., Chappell, J., Kinsley, L., Shelley, M., Murray-Wallace, C.V., Spötl, C., Taylor, L., 2005. In situ U-series dating by laser ablation multi-collector ICPMS: new prospects for Quaternary geochronology. Quatern. Sci. Rev. 24, 2523e2538. Eggins, S.M., Grün, R., Pike, A.W.G., Shelley, A., Taylor, L., 2003. 238U, 232Th profiling and U-series isotope analysis of fossil teeth by laser ablation-ICPMS. Quatern. Sci. Rev. 22, 1373e1382. Feibel, C.S., 2008. Microstratigraphy of the Kibish hominin sites KHS and PHS, Lower Omo Valley, Ethiopia. J. Hum. Evol. 55, 404e408. Fleagle, J.G., Assefa, Z., Brown, F.H., Shea, J.J., 2008. Paleoanthropology of the Kibish Formation, southern Ethiopia: introduction. J. Hum. Evol. 55, 360e365. Grün, R., Aubert, M., Hellstrom, J., Duval, M., 2010. The challenge of direct dating old human fossils. Quatern. Int. 234, 87e93. Grün, R., Aubert, M., Joannes-Boyau, R., Moncel, M.H., 2008. High resolution analysis of uranium and thorium concentrations as well as U-series isotope distributions in a Neanderthal tooth from Payre using laser ablation ICP-MS. Geochim. Cosmochim. Ac. 72, 5278e5290. Grün, R., Huang, P.H., Huang, W., McDermott, F., Stringer, C.B., Thorne, A., Yan, G., 1998. ESR and U-series analyses of teeth from the palaeoanthropological site of Hexian, Anhui Province, China. J. Hum. Evol. 34, 555e564. Grün, R., Maroto, J., Eggins, S., Stringer, C., Robertson, S., Taylor, L., Mortimer, G., McCulloch, M., 2006. ESR and U-series analyses of enamel and dentine fragments of the Banyoles mandible. J. Hum. Evol. 50, 347e358. Grün, R., Stringer, C., 1991. Electron spin resonance dating and the evolution of modern humans. Archaeometry 33, 153e199. Grün, R., Stringer, C., McDermott, F., Nathan, R., Porat, N., Robertson, S., Taylor, L., Mortimer, G., Eggins, S., McCulloch, M., 2005. U-series and ESR analyses of bones and teeth relating to the human burials from Skhul. J. Hum. Evol. 49, 316e334. Gunz, P., Bookstein, F.L., Mitteroecker, P., Stadlmayr, A., Seidler, H., Weber, G.W., 2009. Early modern human diversity suggests subdivided population structure and a complex out-of-Africa scenario. Proc. Natl. Acad. Sci. 106, 6094e6098. Hellstrom, J., 2006. UeTh dating of speleothems with high initial 230Th using stratigraphical constraint. Quatern. Geochronol. 1, 289e295. Klein, R.G., 2009. The Human Career, third ed. University of Chicago Press, Chicago. Leakey, R.E.F., 1969. Early Homo sapiens remains from the Omo River region of south-west Ethiopia. Nature 222, 1132e1133. McDougall, I., Brown, F.H., Fleagle, J.G., 2005. Stratigraphic placement and age of modern humans from Kibish, Ethiopia. Nature 433, 733e736. McDougall, I., Brown, F.H., Fleagle, J.G., 2008. Sapropels and the age of hominins Omo I and Omo II, Kibish, Ethiopia. J. Hum. Evol. 55, 409e420. Millard, A.R., 1993. Diagenesis of Archaeological Bone: The Case of Uranium Uptake. Ph.D. Dissertation, University of Oxford. Millard, A.R., 2008. A critique of the chronometric evidence for hominid fossils: I. Africa and the Near East 500e50 ka. J. Hum. Evol. 54, 848e874. Millard, A.R., Hedges, R.E.M., 1996. A diffusion-adsorption model of uranium uptake by archaeological bone. Geochim. Cosmochim. Ac. 60, 2139e2152. Pearson, O.M., Fleagle, J.G., Grine, F.E., Royer, D.F., 2008a. Further new hominin fossils from the Kibish Formation, southwestern Ethiopia. J. Hum. Evol. 55, 444e447.
710
M. Aubert et al. / Journal of Human Evolution 63 (2012) 704e710
Pearson, O.M., Royer, D.F., Fleagle, J.G., Grine, F.E., 2008b. A description of the Omo I postcranial skeleton, including newly discovered fossils. J. Hum. Evol. 55, 421e437. Pike, A.W.G., 2000. Uranium Series Dating of Archaeological Bone by Thermal Ionization Mass Spectrometry. Ph.D. Dissertation, University of Oxford. Pike, A.W.G., Hedges, R.E.M., Van Calsteren, P., 2002. U-series dating of bone using the diffusion-adsorption model. Geochim. Cosmochim. Ac. 66, 4273e4286. Sambridge, M., Grün, R., Eggins, S., 2012. U-series dating of bone in an open system: the diffusion-adsorption-decay model. Quatern. Geochronol. 9, 42e53.
Smith, F.H., 1992. Models and realities in modern human origins: the African fossil evidence. Phil. Trans. R. Soc. Lond. B 337, 243e250. Stringer, C., 2007. The origin and dispersal of Homo sapiens: our current state of knowledge. In: Mellars, P., Boyle, K., Bar-Yosef, O., Stringer, C. (Eds.), Rethinking the Human Revolution. McDonald Institute for Archaeological Research Monograph Series, Cambridge, pp. 15e20. Stringer, C., 2012. Lone Survivors: How We Came to Be the Only Humans on Earth. Times Books, New York.
Contents lists available at SciVerse ScienceDirect
Journal of Human Evolution journal homepage: www.elsevier.com/locate/jhevol
Confirmation of a late middle Pleistocene age for the Omo Kibish 1 cranium by direct uranium-series dating Maxime Aubert a, *,1, Alistair W.G. Pike b, Chris Stringer c, Antonis Bartsiokas d, Les Kinsley a, Stephen Eggins a, Michael Day d, Rainer Grün a a
Research School of Earth Sciences, The Australian National University, Canberra ACT 0200, Australia Department of Archaeology and Anthropology, University of Bristol, 43 Woodland Road, Bristol BS8 1UU, UK Department of Palaeontology, The Natural History Museum, London SW7 5BD, UK d Department of History and Ethnology, Democritus University of Thrace, P.O. Box 217, 69100 Komotini, Greece b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 January 2012 Accepted 14 July 2012 Available online 5 September 2012
While it is generally accepted that modern humans evolved in Africa, the specific physical evidence for that origin remains disputed. The modern-looking Omo 1 skeleton, discovered in the Kibish region of Ethiopia in 1967, was controversially dated at w130 ka (thousands of years ago) by U-series dating on associated Mollusca, and it was not until 2005 that AreAr dating on associated feldspar crystals in pumice clasts provided evidence for an even older age of w195 ka. However, questions continue to be raised about the age and stratigraphic position of this crucial fossil specimen. Here we present direct Useries determinations on the Omo 1 cranium. In spite of significant methodological complications, which are discussed in detail, the results indicate that the human remains do not belong to a later intrusive burial and are the earliest representative of anatomically modern humans. Given the more archaic morphology shown by the apparently contemporaneous Omo 2 calvaria, we suggest that direct U-series dating is applied to this fossil as well, to confirm its age in relation to Omo 1. Ó 2012 Elsevier Ltd. All rights reserved.
Keywords: Oldest anatomically modern humans Homo sapiens Laser ablation Africa Ethiopia
Introduction The three fossil humans found in 1967 in the Kibish Formation of the lower basin of the Omo River, southern Ethiopia, come from different locations and stratigraphic contexts (Butzer, 1969; Butzer et al., 1969; Leakey, 1969). Omo 1, a partial skull and associated skeleton was excavated at site KHS in Member I. Omo 2, a calvaria, was a surface find reportedly at site PHS, 2.5 km away on the other side of the Omo River, but allocated to Member I by correlation (however, see below). Omo 3, consisting only of cranial fragments, was apparently derived from Member III. Omo 1 and 3 represent anatomically modern Homo sapiens, but Omo 2 is more archaic in cranial morphology (Day, 1969; Day and Stringer, 1991). Radiocarbon dates for the upper part of the Kibish Formation suggested that Members IeIII lay beyond the practical limits of the method, and this was consistent with uranium-series age estimates of
* Corresponding author. E-mail address: [email protected] (M. Aubert). 1 Present address: Centre for Archaeological Science, School of Earth and Environmental Sciences, University of Wollongong, Wollongong 2522, Australia. 0047-2484/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jhevol.2012.07.006
130 5 ka (thousands of years ago) on Etheria shells from Member I, correlated by Butzer (1969) with the stratigraphic position of the Omo 1 skeleton. In turn this suggested that the Omo 1 skeleton was potentially the oldest known example of the modern form of Homo sapiens. However, serious questions were subsequently raised about the reliability of these U-series dates on mollusc material (see e.g., Smith, 1992 and following discussion). Renewed excavations at the original locality have produced further hominin fossils, including some from the Omo 1 skeleton, recovered in situ (Fleagle et al., 2008; Pearson et al., 2008a, b), and clarified the location and stratigraphic placement of the Omo 2 find (McDougall et al., 2005, 2008; Brown and Fuller, 2008; Feibel, 2008). Associated dating studies correlated AreAr determinations on feldspar crystals in pumice clasts from Member I with regional fluvial cycles and Mediterranean sapropels, suggesting a revised age of between 172 ka and w195 5 ka for Omo 1 and 2 (McDougall et al., 2005, 2008; Brown et al., 2012). This work was in turn re-evaluated by Millard (2008), who proposed an age somewhat younger than 190 ka, while Klein (2009) raised the possibility that the Omo 1 skeleton was in fact a much younger intrusive burial. Direct dating work on the Omo 1 skeleton is thus of critical importance in confirming or denying that this fossil may
M. Aubert et al. / Journal of Human Evolution 63 (2012) 704e710
represent the oldest known example of anatomically modern H. sapiens. We have applied U-series analysis to gain further insights to the age of the fossil. U-series dating of bones is seriously compromised by the fact that bones can accumulate large amounts of uranium following their deposition into sediments. A range of models have been developed to account for this uranium uptake and provide a basis for open system dating. The diffusion-adsorption (D-A) model was developed by Millard (1993) and Millard and Hedges (1996), and refined by Pike (2000) and Pike et al. (2002). It is based on laboratory experiments and assumes a continuous diffusion of uranium from the outside of a bone or tooth towards the interior, and that the partitioning between the bone and solution (groundwater) and the U concentration in the solution are constant. The bone is treated as a homogeneous medium. Under constant conditions, the cross sections of bones that conform to the D-A diffusion model are expected to have both u-shaped Uconcentration and apparent U-series age profiles, with the apparent ages at the surface being closest to the correct age of the sample. Deviations from such ideal profiles can be explained either by leaching or changes in the U concentration in the solution. The D-A model was recently refined by Sambridge et al. (2012). Their Diffusion-Adsorption-Decay (DAD) model expands the D-A model for diffusion of 234U and its decay during the diffusion process. For a given volume in a bone, the DAD model postulates that 234U is continually resupplied by diffusion. As a result, 234U/238U ratios change little over time, consequently DAD model ages are somewhat older than comparative D-A results. All age results presented here are based on the assumptions underlying the DAD model. Samples We have dated two small (w4 g) parietal fragments (A and B) of the Omo Kibish 1 skull that could not be fitted onto the cranial reconstruction and that were retained in London, first at St. Thomas’s Hospital Medical School, and then the Natural History Museum. Fragment A was initially analysed along a single laser track while fragment B was later analysed along a series of laser track on two different planes. For more details on the analysed surfaces see Results and Discussion below. Experimental Laser ablation elemental U, Th and U-series isotope analyses were carried out at the Australian National University. For the experimental set-up, see Eggins et al. (2003, 2005), and for applications of U-series dating on human fossils, see Grün et al. (2005, 2006, 2008). U and Th concentrations were derived from repeated measurements of the NBS-610 standard, U-isotope ratios from the dentine of a rhinoceros tooth from Hexian (sample 1118, see Grün et al., 1998). The ANU Neptune mass spectrometer has only one ion counter that is required for the measurement of both 230Th and 234U. As a consequence, each track is run twice, first for the estimation of the 230 Th/238U ratio, then for the 234U/238U ratio. When U-series data are derived from spot analysis, two closely adjacent holes are drilled. It turns out that the 234U/238U ratio usually changes little within a bone, so that this analytical approach does not introduce any significant additional errors when compared with laser ablation analyses using multiple ion counters (e.g., Hellstrom, 2006; Grün et al., 2010; Duval et al., 2011). To remove surface contamination, the tracks and spots were first cleaned with the laser using a wider spot size. None of the measurements were affected by detrital Th. 230Th/232Th activity ratios were well above 100, mostly above 1000.
705
Measurements were carried out in 1024 ms cycles. Along the tracks, an apparent U-series age was calculated for each cycle. These had 2-s errors in the 20% range. To reduce the noise of the data scatter, a sliding five point average was applied. The errors of the individual age estimates are not shown. Both samples had large pores, some filled with secondary minerals, the U-concentrations dropped considerably (often to less than 1 ppm) from the average of around 30e50 ppm. The low Uconcentrations lead to very large errors in the calculation of Useries isotope ratios, and consequently the mean ages scatter widely. These regions were excluded in the age profiles that lead to much more coherent results when using a five point sliding average. In previous experiments, we found that there are differences in the measured 230Th/234U ratios in the range of 3% depending on the direction of the laser tracks. The cause for this is at present not resolved. To alleviate any problems relating to the motion of the sample in the ablation cell during isotope measurement, we carried out the spot-analyses, which does not require any sample movement. The holes were drilled with the laser set to 137 mm in diameter for 100 s. Ages were calculated from the integrated data and have individual 2-s errors in the range of 5%e6%. The highest 230 Th/234U ratios of the track analyses are about 3% higher than the corresponding spot results. Age calculations During the alpha decay of 238U, a parent atom experiences a strong recoil. In minerals, this recoil leads to a weakened bond between the remaining atoms and the crystal lattice. When minerals are exposed to weathering, 234U atoms are preferably leached. This process leads to the well-known fact that surface and ground waters have excess 234U over 238U (Cherdyntsev, 1971). This has some major implications for the dating of bones. In closed systems, for example for speleothems or corals, U-series ages are calculated from the measured 230Th/234U and 234U/238U ratios. The decay of excess 234U is taken into account in the age calculation. Bones, however, are an open system for uranium and for a given volume, the 234U/238U ratio remains more or less constant by the continuing re-supply of 234U from the outside. As a result, most bones have rather uniform 234U/238U ratios. Indeed, all 234U/238U ratios of Omo fall into a narrow range of between 1.41 and 1.46. Assuming that the 234U/238U ratios remain constant over time, bone ages are calculated from the present 230Th/234U ratios. For more details see (Sambridge et al., 2012). Results and discussion Initially, sample A was analysed along a single laser track. The apparent U-series ages and U-distributions indicated a rapid Uuptake with a best age estimate of 98þ8 6 ka (Fig. 1). We had submitted a paper containing these results in 2005, but unfortunately this coincided with the publication of the Ar/Ar results of McDougall et al. (2005), who had obtained an age estimate on a tuff from Member 1, underlying the human remains, of 195.8 1.6 ka and an estimate of 103.7 0.9 ka for tuff in the upper part of Member 3, about 50 m above the former sample. Considering that the first result was obtained from the same sedimentary unit in which the human remains were found while the second was derived from a unit that was deposited at least two sedimentary cycles later, McDougall et al. (2005) made a convincing case that the age of Omo 1 should be close to 195 ka. Whilst our results did not contradict the findings of McDougall et al. (2005), they pointed to the basic problem of U-series dating of bones, namely that the results are usually minimum age estimates, and that it is difficult to
706
M. Aubert et al. / Journal of Human Evolution 63 (2012) 704e710
Figure 1. Age and U-concentration along a single laser track from sample A measured in 2005. The apparent U-series ages and U-distributions indicated a rapid U-uptake with a best age estimate of 98þ86 ka.
impossible to estimate by how much the U-series results underestimate the correct age of the sample (Pike et al., 2002). In the game of finding the oldest modern humans, a direct minimum age of 98 ka would have been sensational some 20 years ago (e.g., Grün and Stringer, 1991), but in light of a 195 ka Ar/Ar age on the sediments, the U-series results were dismissed out of hand. Detailed U-series mapping of a Neanderthal tooth from Payre (Grün et al., 2008) indicated that bones and teeth may contain domains with very different apparent U-series ages, the oldest being close to the expected age range. The sample B had a nearly flat surface, presenting a cross section through the bone. To minimise destruction, this surface was evened out with fine sand paper. Six tracks were run across this bone (Fig. 2). The U-concentration profiles along the six tracks are either flat or show some increased concentrations close to the outer surfaces. This is more or less expected from the DiffusionAdsorption model (Pike et al., 2002). However, the apparent Useries data are not entirely symmetrical, the apparent U-series ages at the top of the fragment are somewhat older than at the base. Furthermore, close to the surfaces, the apparent U-series results are often younger, particularly where the U-concentrations increase (e.g., Track 2, between cycles 240 and 260, or Track 5, between cycles 1 and 10). This points to a secondary overprint of uranium. Track 1 generally replicated the earlier results with ages around 100 ka. However, the apparent U-series age estimates gradually increase, particularly in the upper part to the right, where apparent age estimates of around 170 ka are reached along Tracks 5 (around cycle 60) and 6 (between cycles 100e120). This may either point to an earlier U-uptake phase, or to U-leaching. In the scanned surface, no evidence for leaching is apparent, which would be associated with lower U-concentrations. Nevertheless, uranium could have been lost through the exposed surface, i.e., perpendicular to the measurement plane. It was decided to cut the fragment about 2 mm below the original surface with a diamond wire saw. The cutting width was around 110 mm. The subsequent U-series analyses were carried out on the underside of the first plane (Fig. 3). For easier comparison, the image is flipped. The U-concentrations in this second plane are somewhat lower than in the upper plane. Ignoring the pores, the Uconcentration profiles are either flat or increase towards the outer surfaces. Similar to the upper plane, the apparent U-series age estimates indicate secondary overprints in some domains close to the surface (e.g., Tracks 1, 5 and 6, close to the ends of the scans in
Fig. 3). Otherwise, the apparent age profiles are either flat or Ushaped, with a tendency of older ages towards the upper part of the fragment (compare beginnings and ends of the tracks). At the upper right hand side (particularly Tracks 9 and 10), ages of more than 200 ka were obtained. There is no evidence for lateral U-leaching. The general agreement between the apparent U-series estimates in the upper and lower planes indicates that the old ages were not produced by U-leaching along a perpendicular direction. When running the ablation cell, it was observed that some differences in the 230Th/238U ratio may occur depending on the absolute track direction, i.e., when the same track is analysed after being rotated by 90 , the results may be somewhat different. This, to our knowledge, does not affect the relative results along a track, or between parallel tracks, as long as they are measured in the same direction. To address this problem, spot analyses were carried out in the domain that yielded the oldest apparent U-series ages (see dotted square in Fig. 3, and Fig. 4). As mentioned above, a U-series age calculation using spots consists of two separate analyses. Therefore, only about half of the holes can be used for a U-series calculation (see Fig. 4A). However, spot analyses have the advantage of not being dependent on movements within the cell, and are averaged over much longer measuring times, resulting in significantly smaller individual errors. The calculated mean ages are shown in Fig. 4B. The spots provided old ages on the same vicinity where the tracks also yielded the oldest apparent U-series age estimates. As mentioned above, this domain did not show any signs of U-leaching expressed in lower U-concentrations. In addition, the 234U/238U ratios change from around 1.44 to 1.46 close to the outside (upper part of larger square in Fig. 4A) to 1.41 to 1.42 further to the inside. At the same time, the apparent 230Th/234U ages decrease from around 160 to 180 ka close to the outside to 140 to 150 ka further inside. This is consistent with the predictions of the DAD model (Sambridge et al., 2012), i.e., a function of U-uptake. When Uleaching occurs, the 234U/238U ratios in the affected volumes generally decrease due to preferential leaching of 234U (e.g., Duval et al., 2011). The average ages are younger than those observed along the tracks, but are based on significantly larger volumes. A detailed 3D study of U concentrations and U-series isotope ratios in an early Pleistocene horse tooth (Duval et al., 2011) showed that small-scale uranium redistribution may lead to significant changes in the respective 230Th/238U ratios. This may well explain some of the high apparent ages in the inner square, or the low age of 92 ka at the base between tracks 10 and 11. Averaging over larger areas will address the problem of micro-redistribution. Concentrating at the end of tracks 9 to 11 (inner square in Fig. 4), an average age of 171 16 ka is obtained, averaging a much larger area (outer square in Fig. 4) results in 162 14 ka. The much larger errors of the averages compared to the errors of the individual results are an indication of U-mobilisation. Assuming that Omo1 has indeed an age of close to 195 ka (McDougall et al., 2005), these older domains in the bone approximate this age closely. Our detailed U-series analyses also point to the deficiency of Useries dating of bones: if the analysis is confined to a small section, e.g., a single track, severe underestimations may occur. These underestimations are probably due to changes in the U-concentrations in the percolating waters or mineralogical changes in the bones that allow the adsorption of higher U-concentrations with time. Furthermore, many bone sections may well not contain any domains that have preserved the original isotopic signature. As a result, most U-series age estimations on bones have to be regarded as minimum age estimates, even when D-A or DAD model predictions are apparently fulfilled (e.g., Fig. 1). On the other hand, data sets complying with the D-A or DAD models will not overestimate the age of a bone.
M. Aubert et al. / Journal of Human Evolution 63 (2012) 704e710
707
Figure 2. Age and U-concentration along individual tracks 1e6 from sample B (upper plane). The surface was obtained by slightly abrading the original surface with fine sand paper. Track 1 generally replicated the earlier results from sample A with ages around 100 ka. However, the apparent U-series age estimates gradually increase, particularly in the upper part to the right, where apparent age estimates of around 170 ka are reached along Tracks 5 (around cycle 60) and 6 (between cycles 100e120). The analysed surface area measures approximately 30 mm by 15 mm.
708
M. Aubert et al. / Journal of Human Evolution 63 (2012) 704e710
Figure 3. Age and U-concentration along individual tracks 1e12 from sample B (lower plane). The surface was obtained by cutting the fragment about 2 mm below the original surface with a diamond wire saw. The subsequent U-series analyses were carried out on the underside of the first plane. Similar to the upper plane, the apparent U-series age estimates gradually increase towards the upper part of the fragment. At the upper right hand side (particularly Tracks 9 and 10), ages of more than 200 ka were obtained. The general agreement between the apparent U-series estimates in the upper and lower planes indicates that the old ages were not produced by U-leaching along a perpendicular direction. The analysed surface area measures approximately 30 mm by 15 mm.
M. Aubert et al. / Journal of Human Evolution 63 (2012) 704e710
709
Acknowledgements Aspects of this study were supported by Australian Research Council Grants DP0664144 (Grün et al.) ‘Microanalysis of human fossils: new insights into age, diet and migration’ and DP0666084 (Roberts et al.) ‘Out of Africa and into Australia: robust chronologies for turning points in modern human evolution and dispersal’. Chris Stringer is a member of the Ancient Human Occupation of Britain project, funded by the Leverhulme Trust, and his research is supported by the Calleva Foundation. References
Figure 4. Results of drilling holes with the laser into the domain that yielded the oldest apparent U-series ages (see dotted square in Fig. 3). A: 234U/238U and 230Th/238U ratios. B: Age estimates. Note that the average ages are somewhat younger than those observed along the tracks, but are based on significantly larger volumes.
Implications The results presented here illustrate that a wide range of apparent U-series ages can be generated on a particular sample and that careful evaluation of the data is required. The younger apparent ages obtained here appear to be related to higher Uconcentrations pointing out to a later overprint of uranium. This is further illustrated by the higher U-concentrations and younger age estimates in the upper plane (sample B) compared with the lower plane. It also appears that the domain that yielded the oldest apparent U-series ages was not produced by U-leaching so that the minimum age estimate for this sample is w155e187 ka. In spite of the methodological complications involved in Useries analysis of bones, we expect that laser ablation U-series dating will lead to further critical insights into the timing of modern human evolution, and we recommend that this technique should be applied to the morphologically distinct Omo 2 calvaria to test the stratigraphic and chronological relationship between the Omo 1 and Omo 2 fossils. If their contemporaneity can be established, this will further demonstrate the great morphological variability displayed among early H. sapiens fossils in Africa, and suggestions of more complex scenarios for the evolutionary origins of our species in that continent (Stringer, 2007, 2012; Gunz et al., 2009). Notwithstanding those issues, our new direct dating of the Omo 1 cranium seem to confirm its middle Pleistocene age within Member I of the Kibish Formation, and refute any suggestion that it represents a much later intrusive burial. Presently, the Omo 1 specimen is the earliest known representative of the anatomically modern skeletal pattern.
Brown, F.H., Fuller, C.R., 2008. Stratigraphy and tephra of the Kibish Formation, southwestern Ethiopia. J. Hum. Evol. 55, 366e403. Brown, F.H., McDougall, I., Fleagle, J.G., 2012. Correlation of the KHS Tuff of the Kibish Formation to volcanic ash layers at other sites, and the age of early Homo sapiens (Omo I and Omo II). J. Hum. Evol. 63 (4), 577e585. Butzer, K.W., 1969. Geological interpretations of two Pleistocene hominid sites in the Lower Omo Basin. Nature 222, 1133e1135. Butzer, K.W., Brown, F.H., Thurber, D.L., 1969. Horizontal sediments of the lower Omo Valley: the Kibish Formation. Quaternaria 11, 15e29. Cherdyntsev, V.V., 1971. Uranium-234. Israel Program for Scientific Translations, Jerusalem. Day, M.H., 1969. Omo human skeletal remains. Nature 222, 1135e1138. Day, M.H., Stringer, C.B., 1991. Les restes crâniens d’Omo Kibish et leur classification à l’intérieur du genre Homo. L’Anthropologie 94, 573e594. Duval, M., Aubert, M., Hellstrom, J., Grün, R., 2011. High resolution, three dimensional mapping of U-concentrations and U-series isotopes in an Early Pleistocene equid tooth from Fuente Nueva-3 (Orce, Andalusia, Spain) using laser ablation ICP-MS. Quartern. Geochronol. 6, 458e467. Eggins, S.M., Grün, R., McCulloch, M.T., Pike, A.W.G., Chappell, J., Kinsley, L., Shelley, M., Murray-Wallace, C.V., Spötl, C., Taylor, L., 2005. In situ U-series dating by laser ablation multi-collector ICPMS: new prospects for Quaternary geochronology. Quatern. Sci. Rev. 24, 2523e2538. Eggins, S.M., Grün, R., Pike, A.W.G., Shelley, A., Taylor, L., 2003. 238U, 232Th profiling and U-series isotope analysis of fossil teeth by laser ablation-ICPMS. Quatern. Sci. Rev. 22, 1373e1382. Feibel, C.S., 2008. Microstratigraphy of the Kibish hominin sites KHS and PHS, Lower Omo Valley, Ethiopia. J. Hum. Evol. 55, 404e408. Fleagle, J.G., Assefa, Z., Brown, F.H., Shea, J.J., 2008. Paleoanthropology of the Kibish Formation, southern Ethiopia: introduction. J. Hum. Evol. 55, 360e365. Grün, R., Aubert, M., Hellstrom, J., Duval, M., 2010. The challenge of direct dating old human fossils. Quatern. Int. 234, 87e93. Grün, R., Aubert, M., Joannes-Boyau, R., Moncel, M.H., 2008. High resolution analysis of uranium and thorium concentrations as well as U-series isotope distributions in a Neanderthal tooth from Payre using laser ablation ICP-MS. Geochim. Cosmochim. Ac. 72, 5278e5290. Grün, R., Huang, P.H., Huang, W., McDermott, F., Stringer, C.B., Thorne, A., Yan, G., 1998. ESR and U-series analyses of teeth from the palaeoanthropological site of Hexian, Anhui Province, China. J. Hum. Evol. 34, 555e564. Grün, R., Maroto, J., Eggins, S., Stringer, C., Robertson, S., Taylor, L., Mortimer, G., McCulloch, M., 2006. ESR and U-series analyses of enamel and dentine fragments of the Banyoles mandible. J. Hum. Evol. 50, 347e358. Grün, R., Stringer, C., 1991. Electron spin resonance dating and the evolution of modern humans. Archaeometry 33, 153e199. Grün, R., Stringer, C., McDermott, F., Nathan, R., Porat, N., Robertson, S., Taylor, L., Mortimer, G., Eggins, S., McCulloch, M., 2005. U-series and ESR analyses of bones and teeth relating to the human burials from Skhul. J. Hum. Evol. 49, 316e334. Gunz, P., Bookstein, F.L., Mitteroecker, P., Stadlmayr, A., Seidler, H., Weber, G.W., 2009. Early modern human diversity suggests subdivided population structure and a complex out-of-Africa scenario. Proc. Natl. Acad. Sci. 106, 6094e6098. Hellstrom, J., 2006. UeTh dating of speleothems with high initial 230Th using stratigraphical constraint. Quatern. Geochronol. 1, 289e295. Klein, R.G., 2009. The Human Career, third ed. University of Chicago Press, Chicago. Leakey, R.E.F., 1969. Early Homo sapiens remains from the Omo River region of south-west Ethiopia. Nature 222, 1132e1133. McDougall, I., Brown, F.H., Fleagle, J.G., 2005. Stratigraphic placement and age of modern humans from Kibish, Ethiopia. Nature 433, 733e736. McDougall, I., Brown, F.H., Fleagle, J.G., 2008. Sapropels and the age of hominins Omo I and Omo II, Kibish, Ethiopia. J. Hum. Evol. 55, 409e420. Millard, A.R., 1993. Diagenesis of Archaeological Bone: The Case of Uranium Uptake. Ph.D. Dissertation, University of Oxford. Millard, A.R., 2008. A critique of the chronometric evidence for hominid fossils: I. Africa and the Near East 500e50 ka. J. Hum. Evol. 54, 848e874. Millard, A.R., Hedges, R.E.M., 1996. A diffusion-adsorption model of uranium uptake by archaeological bone. Geochim. Cosmochim. Ac. 60, 2139e2152. Pearson, O.M., Fleagle, J.G., Grine, F.E., Royer, D.F., 2008a. Further new hominin fossils from the Kibish Formation, southwestern Ethiopia. J. Hum. Evol. 55, 444e447.
710
M. Aubert et al. / Journal of Human Evolution 63 (2012) 704e710
Pearson, O.M., Royer, D.F., Fleagle, J.G., Grine, F.E., 2008b. A description of the Omo I postcranial skeleton, including newly discovered fossils. J. Hum. Evol. 55, 421e437. Pike, A.W.G., 2000. Uranium Series Dating of Archaeological Bone by Thermal Ionization Mass Spectrometry. Ph.D. Dissertation, University of Oxford. Pike, A.W.G., Hedges, R.E.M., Van Calsteren, P., 2002. U-series dating of bone using the diffusion-adsorption model. Geochim. Cosmochim. Ac. 66, 4273e4286. Sambridge, M., Grün, R., Eggins, S., 2012. U-series dating of bone in an open system: the diffusion-adsorption-decay model. Quatern. Geochronol. 9, 42e53.
Smith, F.H., 1992. Models and realities in modern human origins: the African fossil evidence. Phil. Trans. R. Soc. Lond. B 337, 243e250. Stringer, C., 2007. The origin and dispersal of Homo sapiens: our current state of knowledge. In: Mellars, P., Boyle, K., Bar-Yosef, O., Stringer, C. (Eds.), Rethinking the Human Revolution. McDonald Institute for Archaeological Research Monograph Series, Cambridge, pp. 15e20. Stringer, C., 2012. Lone Survivors: How We Came to Be the Only Humans on Earth. Times Books, New York.