sign in sign up
<<
Home , Bodo cranium, Middle Awash

African erectus: Old Radiometric Ages and Young Assemblages in the Middle Awash Valley, Author(s): J. D. Clark, J. de Heinzelin, K. D. Schick, W. K. Hart, T. D. White, G. WoldeGabriel, R. C. Walter, G. Suwa, B. Asfaw, E. Vrba and Y. H.-Selassie Reviewed work(s): Source: Science, New Series, Vol. 264, No. 5167 (Jun. 24, 1994), pp. 1907-1910 Published by: American Association for the Advancement of Science Stable URL: http://www.jstor.org/stable/2883979 . Accessed: 30/12/2012 10:38

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

.

American Association for the Advancement of Science is collaborating with JSTOR to digitize, preserve and extend access to Science.

http://www.jstor.org

This content downloaded on Sun, 30 Dec 2012 10:38:51 AM All use subject to JSTOR Terms and Conditions REPORTS

REFERENCESAND NOTES Honda, D. A. Yuen, S. Balachandar, D. Reuteler, J. Earth Sci. 30, 881 (1993). Science 259, 1308 (1993)], a number of errors in 15. C. W. Wicks Jr. and M. A. Richards, Science 261, 1 G. F. Davies, J. Geophys. Res. 89,6017 (1984); K. the implementation of phase effects, including the 1424 (1993). C. Creager and T. H. Jordan, ibid. 91, 3573 limiting of the phase-related body forces to the 16. G. T. Jarvis and D. P. McKenzie, J. Fluid Mech. (1986); J. Phipps Morgan and P. M. Shearer, phase boundary, prevented significant asymmet- 96, 515 (1980). Nature 365, 506 (1993). ric phase effects from being seen. 17. S. A. Weinstein, P. L. Olson, D. A. Yuen, Geophys. 2. D. P. McKenzie, J. M. Roberts, N. 0. Weiss, J. 7. R. Bohler and A. Chopelas, Geophys. Res. Lett. Astrophys. Fluid Dyn. 47, 157 (1989). Fluid Mech. 62, 465 (1974); F. M. Richter and D. 18, 1147 (1991). 18. W. L. Zhao, D. A. Yuen, S. Honda, Phys. Earth P. McKenzie, J. Geophys. 44, 441 (1978). 8. T. Kato and M. Kumazawa, in Dynamic Processes Planet. Inter. 72, 185 (1992). of Material Transformation in the Earth's 3. C. R. Bina, Rev. Geophys. (suppl.) 29 (1991). Interior, 19. E. Ito and E. Takahashi, J. Geophys. Res. 94, M. Marumo, Ed. (Terra, Tokyo, 1990), pp. 277- 10637 4. D. L. Anderson, J. Geophys. Res. 92, 13968 (1989). 208. 20. S. Honda and D. A. Yuen, Earth Planet. (1987). Sci. Lett. 9. Y. Fei, S. K. Saxena, A. Navrotsky, J. Geophys. 96, 349 (1990). 5. Other phase effects include latent heat that is Res. 95, 6915 (1990). 21. P. J. Tackley, D. J. Stevenson, G. A. Glatzmaier, either absorbed (endothermic) or released (exo- 10. A. E. Ringwood, in The Earth, Its Origin, Structure G. Schubert, Nature 361, 699 (1993). thermic) during phase transitions. The effects of and Evolution, M. W. McElhinny, Ed. (Academic 22. D. L. Anderson and J. Bass, ibid. 320, 321 (1986); latent heat are generally in the sense opposite to Press, New York, 1979), pp. 1-58. R. Jeanloz and E. Knittle, Philos. Trans. R. Soc. that of the local body forces associated with the 11. M. Liu and C. G. Chase, J. Geophys. Res. 94, London Ser. A 328, 417 (1989). distortion of phase boundaries Schubert, D. A. [G. 5571 (1989); M. McNutt, Geophys. Monogr. Am. 23. B. H. Hager and M. A. Richards, Philos. Trans. R. Yuen, D. L. Turcotte, Geophys. J. R. Astron. Soc. Geophys. Union43, 1 (1987). Soc. London Ser. A 328, 309 (1989). 42, 705 (1975)]. However, the dynamic effects of 12. U. R. Christensen and D. A. Yuen, J. Geophys. 24. Supported by the Research Board of the Univer- latent heat and other phase effects are usually Res. 90, 10291 (1985).' sity of Missouri. I thank C. Chase, J. Engeln, and much smaller than the buoyancy force induced by 13. M. Liu, D. A. Yuen, W. Zhao, S. Honda, Science P. Nabelek for reading an early version of the phase-boundary distortion (12). 252, 1836 (1991). manuscript. 6. Although multiple phase boundaries were consid- 14. P. Machetel and P. Weber, Nature 350, 55 ered in (13) and some follow-up studies [(18); S. (1991); L. P. Solheim and W. R. Peltier, Can. 8 February 1994; accepted 10 May 1994

AfricanHomo erectus: Old RadiometricAges uted in this area. We do not follow a previousstratigraphic framework formulat- and YoungOldowan Assemblages in the ed for the MiddleAwash (6, 7) becauseof MiddleAwash Valley,Ethiopia the difficulty in correlating sedimentary units acrossa broadregion markedby con- siderablefaulting and lackingprecise chro- J. D. Clark,*J. de Heinzelin, K. D. Schick, W. K. Hart, nometric control. Our fieldworkconcen- T. D. White, G. WoldeGabriel,R. C. Walter, G. Suwa, B. Asfaw, tratedon MiddlePleistocene archaeological E. Vrba, Y. H.-Selassie and paleontologicaloccurrences extending across the modem Bodo, Dawaitoli, and Fossils and artifacts recovered from the middle Awash Valley of Ethiopia's Afar depression Hargufiacatchments (Fig. 1). These are sample the Middle transition from to Homo sapiens. Ar/Arages, incorporatedin fluvial sediments derived biostratigraphy, and tephrachronology from this area indicate that the Pleistocene Bodo fromwadis or seasonalrivers on the eastern hominid cranium and newer specimens are approximately 0.6 million years old. Only side of the basinand a majorriver along the Oldowan chopper and flake assemblages are present in the lower stratigraphic units, but basinaxis. These MiddlePleistocene depos- bifacial artifacts are consistently prevalent and widespread in directly overlying its are in fault contact with a sequence of deposits. This technological transition is related to a shift in sedimentary regime, supporting LowerPleistocene and depositsto the hypothesis that Middle Pleistocene -Oldowan assemblages represent a behavioral the east. The Middle Pleistocene deposits facies of the Acheulean industrial complex. are succeededby a formationexposed as a fault block to the west that bears typical Middle Stone Age artifacts provisionally attributedto the Late Pleistocene. A Homo cranium was found at Bodo, Acheulean artifacts has been widely as- Most of the in situ Middle Pleistocene MiddleAwash Valley, in Ethiopia'sAfar in sumedbased on theirmorphologies (2), but depositsin the Bodo-Hargufiaarea are in a 1976 (1). An age of-3-50,000 yearsfor this in the absence of radiometricdating. A sedimentaryunit now informallydesignated fossil and stratigraphically associated second hominid's parietal was found in "u" (Fig. 1). This -35-m-thick package 1981 (BOD-VP-1/1) (3, 4) and a distal comprisesfour cyclical sedimentaryunits J. D. Clark, T. D. White, Y. H.-Selassie, Laboratory for humerusfragment was recoveredin 1990 and is dividedinto subunitsul to u4 (u4 is Evolutionary Studies, Department of Anthro- (BOD-VP-1/2). These specimens straddle typicallytruncated by later erosion). Each pology, University of California, Berkeley, CA 94720, the traditionalmorphological interface be- subunitconsists of gravel sand to clay silt USA. tween J. de Heinzelin, Institut Royal des Sciences Naturelles Homo erectusand Homo sapiens-a with calcic soils and concretionsat the top. de Belgique, 1040 Brussels, Belgium. transitionwhose age is poorlydefined. The Sand and silt of ul and u2 are indicativeof K. D. Schick, Department of Anthropology, Indiana Bodo craniumexhibits cut marksindicating deposition in a stabilized alluvial plain. University, Bloomington, IN 47405, USA. defleshing (5). This specimen is W. K. Hart, Department of Geology, Miami University, usually Those of u3 and u4 reflect increasingly Oxford, OH 45056, USA. referredto as "archaic"Homo sapiens,as shifting silty overbankand sandy channel G. WoldeGabriel, EES-1/D462, Los Alamos National have other inadequatelydated specimens deposits. A normal fault (fault 6) to the Laboratory, Los Alamos, NM 87545, USA. from Europe (Arago, Petralona)and Asia R. C. Walter, Geochronology Center, Institute of Hu- west separates ul to u4 from a 15- to man Origins, Berkeley, CA 94709, USA. (Yunxian). 20-m-thick sequence of deposits that ap- G. Suwa, Department of Anthropology, University of We recently recoveredadditional Mid- pearsto representthe continuationof unit Tokyo, 113 Tokyo, Japan. dle Pleistocene vertebratefossils at Bodo. u. We informallydesignate the latter B. Asfaw, Post Office Box 5717, Addis Ababa, Ethio- beds pia. In this report, we establish chronostrati- as unit u-t. These beds contain abundant E. Vrba, Department of Geology and Geophysics, Yale graphic control for deposits at Bodo and Middle Pleistocene vertebrate fossils in- University, New Haven, CT 0651 1, USA. elucidate the relations between Oldowan cluding the three fossil hominidsfrom the *To whom correspondence should be addressed. and Acheulean assemblageswidely distrib- upper Bodo sand unit (UBSU) (1). They

SCIENCE * VOL. 264 * 24 JUNE 1994 1907

This content downloaded on Sun, 30 Dec 2012 10:38:51 AM All use subject to JSTOR Terms and Conditions also include a number of occurrencesof N A vAas ier Acheuleanartifacts. The UBSU is separat- A - Acheuleanstone tool location (A#) * Hargufia s, ed from the top of ul by at least four * 4 Oldowanstone tool location(A#) eDawaitoli sedimentarycycles of unknownduration. o ?- Fossilvertebrate location (VP#) * Bodo We sampledtwo differentvitric tephra _u0 Volcanichorizon (MA prefix) horizonswithin units u and u-t for Ar/Ar \-Geochemicalcorrelation ;<-;i ...I... Sandand unit8< tipaa>>Sg dating (Fig. 1). One of these [MA90-20 gravel (Hargufia) = MA90-7 (Dawaitoli) = u-t MA90-23 (Bodo)] separatesul and u2 and Bodocranium; is correlatedon the basis of stratigraphic BOD-VP-1/1;A, - BOD-All1 Fault6 Dawaitoli Hargufia position, similaritiesin glass morphology = BOD-VP-1/2 u4 _ R -u-BO3D-A2,8 w u3 AHAR-A3 (up.), and phenocrystcontent, and geochemistry o~~~~~~~~~~AI A4 This correlationis emphasized 9-4 A O DAW-A5 Ao HAR-A2,A3 (Table 1). 0 6 (up.),9 A 0 (low.) by the physicaland chemicalcharacteristics DAW-A7,8 other u-t Bodo u2 u2 of this horizonas comparedto spatial- OAW-A6 u ly and temporallyassociated tephra. Sam- _ 0 v (low.) v 90-23 ********* 90-7 ****** 90-20 Ar/ArAge= ples were dated by the laser-fusion40Ar/ 0.64 +-0.03 Ma 39Ar method [following methods described ulb -4- BOD-A3E in (8)]. No volcanic horizonswere found 0 - above the UBSU. We attempted to date o +BOD-A4 90-6 ul tuffMA90-24 from u-t, immediatelybelow ula -4-DAW-Al1 the UBSU, but this ash was too contam- *o inatedby detritalfeldspars. The datedsam- 3.5km i 2.5km ples MA90-20 and MA90-23 contain small Bodo Dawaitoli Hargufia South Distance between sections North (<0.10 to 0.25 mm) rare potassiumfeld- spars.Although the laser-fusionmethod is Fig. 1. Stratigraphic relations of Middle Pleistocene deposits (units u and u-t) in the Bodo, Dawaitoli, generallywell suitedfor the datingof small and Hargufia areas. These deposits are in fault contact with Lower Pleistocene deposits to the east grains, the combination of the fine grain and Upper Pleistocene deposits to the west. Subunit ul begins with disconformable thick lenses of size and young age of these feldsparsresults coarse gravels resulting from the readjustment of the topography to tectonics. Subunit u2 is in significantscatter in the data. regularly cyclothemic, with its lower part including a widespread volcanic marker tuff horizon. Feldsparsfrom MA90-20 (-0.5 mm) Peculiar stromatolithic structures occur locally in the middle part of the subunit. Subunit u3 is a yielded six single grain ages, rangingfrom variable thickness of valley fills, from fossiliferous sands to silts and clay-silts. Subunit u4 is usually truncated or absent. Sedimentological, structural, archaeological, and biochronological consider- 0.51 million years ago (Ma) to 0.73 Ma, ations suggest that sequence u-t was emplaced atop unit u before fault 6. Unit u-t and fault 6 are and two multiplegrain ages of 0.39 Ma and illustrated here only at Bodo, although the same structural relations separate units u and u-t at 0.57 Ma. Combined, these data yielded a Dawaitoli and Hargufia. weighted mean age of 0.55 + 0.03 Ma (Table 2). A third multiple grain analysis yielded an age of 1.68 Ma, indicating the lyticaluncertainty inherent in the measure- with the Middle Awash radiometricdates. presenceof detritalfeldspars. Feldspars from ment of such small, young grains, we be- A distal third of a hominid humerus MA90-23 are much more finely grained lieve that the best estimate for the age of lackingan articularend (BOD-VP-1/2)was (<0.25 mm), and we fused three to five this horizon is suggested by the pooled recoveredin 1990 from the surfaceof the grainsto increasethe precisionof the mea- resultsof 0.64 + 0.03 Ma. However, the upper Bodo sands -100 m north of the surement. Five multiple grain ages range single grain ages from MA90-20 suggesta original Bodo cranium discovery. Aside from0.69 to 0.87 Ma and yield a weighted minimumage of 0.55 ? 0.03 Ma for the froma greatercortical thickness, this fossil mean age of 0.74 ? 0.03 Ma. (Table 2). base of subunit u2. Overall, these results has no morphologicalfeatures noticeably The MA90-23 feldsparsare compositionally indicate a temporalcorrelation with upper differentfrom modernhuman comparative heterogeneousand are consistently older membersof the Olorgesailieformation in specimens.This new Bodo humerusis ap- than feldsparsfrom MA90-20. These differ- Kenya(Olorgesailie members 10 and 14 are preciablysmaller than many modem hu- ences may be due to contaminationin the 0.66 Ma and 0.49 Ma, respectively)(9). man humeri.The two other Bodo hominid MA90-23 multiplegrain analyses. The archaeological technologies and individuals are representedby large and Pooledtogether and withoutthe obvious faunal assemblagesfrom units u and u-t in robustcranial remainsrecovered from the contaminant(L# 4980-09) (Table 2), the the Middle Awash are currentlyindistin- same locality. The relatively small size of MA90-20 and MA90-23 analysesyield a guishableand areeffectively contemporane- the new humerusmay reflect pronounced weighted mean age of 0.64 + 0.03 Ma ous with similarcollections from members sexual dimorphismin this hominid form, (Table 2). Such a pooling is justifiedbe- 10 and 14 of the Olorgesailieformation. In also suggested by broadly contemporary, cause an age probabilityspectrum depicts a additionto the Bodo hominids,the paleon- smallercranial remains from other African normal, Gaussian distribution indicating tologicalsample from the overlyingu-t beds sites such as Sale and Ndutu. Severalwork- that these two samplesbelong to a single at Bodo yielded excellent cranial material ers have identifiedprimitive charactersof age population. Furthermore,an isochron of Kolpochoerusmajus and Alcelaphusbusela- the Bodo crania that accord well with an plot of the pooleddata yields an age of 0.64 phus, Damaliscusniro, and Syncerusacoelo- older age for the fossils than previously ? 0.04 Ma, an initial 40Ar/36Arintercept tus. The Theropithecusoswaldi and K. majus thought (1, 3). However, most workers of 300 ? 17, and a mean standardweight from u and u-t at Bodo (10), Dawaitoli, have identifiedthe Bodo specimen as "ar- difference value of 1.6. Taken together, and Hargufiaindicate a good biochrono- chaic" Homo sapienson the basis of its these results indicate that the pooled data logical match for Olduvai bed IV and the derived morphologycompared to typical fit the criteria for a single age population, upper levels at Olorgesailie. The Bodo Homoerectus of Java and China. with no systematic errors. Given the ana- bovids fit this assessmentand are in accord Our placementof the Bodo craniumat

1908 SCIENCE * VOL. 264 * 24 JUNE 1994

This content downloaded on Sun, 30 Dec 2012 10:38:51 AM All use subject to JSTOR Terms and Conditions X REPORTS

Table 1. Compositional data for Hargufia-Dawaitoli-Bodo tuffs. See Fig. 1 for geographic and draulicagency to accountfor this observation stratigraphic locations. All values are given in parts per million except Fe2O3, A1203, and CaO, which (23). The MiddleAwash data are consistent are given as weight percent (32). with such interpretations.Future use of the landscapeapproach to interpretarchaeologi- Location Fe2O3 A1203 CaO Ba Mn Nb Rb Sr Ti Y Zn Zr cal assemblages(24, 25) will explore late MA90-7 6.06 11.51 0.52 985 1646 119 116 23 2177 99 232 676 Oldowanassemblages as activityfacies of the MA90-20 6.09 11.62 0.53 985 1701 125 116 18 2241 96 219 659 Acheulean(26). MA90-23 6.11 11.51 0.56 1001 1743 124 116 21 2278 98 223 663 Olderradiometric ages (0.64 Ma) for the MA90-6 7.11 10.46 0.43 572 2031 135 100 6 2319 102 256 831 MiddleAwash and for upper Olorgesailie (0.6 MA90-24 8.79 7.98 0.35 1013 2284 138 96 46 1480 119 301 681 to 0.74 Ma) (9) combinewith the discovery of much older (1.0 to 1.5 Ma) Acheulean occurrencesat Konso-Gardula(16), Kesem- Table 2. Laser-fusion 40Ar/39Ar results. The (-0.5 Ma rather than 0.3 Ma). Kebena(27), andBouri (28) to emphasizethe laboratory run number is indicated by L#, the Artifactswere excavated at HAR-A4 and long durationof the Acheulean.These find- number of grains measured in each fusion is DAW-A6 and found erodingfrom individ- ings couple with new evidence for possible indicated by n, and the percent radiation is indicated by % Rad. (33). The average and SD ual horizonsat variouslocalities in the three early(1.6 to 1.8 Ma) hominidoccupation of of the Ca/U ratio as well as the weighted mean modemdrainages. Assemblages in ul andu2 Java(29) to showthat the Old Worldorigins, and uncertainty, respectively, are 0.0133, consist of simplecores and flakestypical of dispersal,culture, and evolution of Horno 0.0127, 0.554, and 0.032 for MA90-20 (J = Oldowan technology; bifaces are notably erectuswere very complex. 0.000315 ? 0.000002); 0.2819, 0.4528, 0.739, absent.Flakes are predominantlycortical or and 0.028 for MA90-23 (J = 0.0003187 ? plain-platformed, primarily unmodified, REFERENCESAND NOTES 0.000001); and 0.1166, 0.2949, 0.639, and retouch. Five pairs of 0.034 in the pooled results (L#4980-01 to with rare, minimal 1. G. C. Conroy, C. J. Jolly, D. Cramer,J. Kalb, L#4980-08 for MA90-20 and L#5005-01 to conjoiningflakes were found at BOD-A3. In Nature275, 67 (1978). L#5005-05 for MA90-23). The material was contrast,Acheulean artifactsdominated by 2. G. Brauerand E. J. Mbua,J. Hum. Evol.22, 79 sanidine except for L#5005-03 (Plagioclase) relativelywell made, bifacialhand axes and (1992). and L#5005-05 (anorthoclase). 3. B. Asfaw,Am. J. Phys. Anthropol.61, 367 (1983). cleaversfirst become abundant in subunitu3 4. J. D. Clarket al., Nature307, 423 (1984). and above, consistentlyfound eroding from 5. T. White,Am. J. Phys. Anthropol.69, 503 (1986). 6. J. E. Kalb et al., Nature298, 17 (1982). L# n Ca/K R/ Age (Ma) ? Ma alluvial sedimentsassociated with shallow 7. J. E. Kalb,E. B. Oswald,A. Mebrate,S. Tebedge, channels or channel banks. Oldowan and C. J. Jolly, Newsl. Stratigr.11, 95 (1982). MA90-20 Acheulean artifactsare mostly in a fresh, 8. R. C. Walter,P. C. Manega,R. L. Hay,R. E. Drake, 4980-01 1 0.0122 33.8 0.533 0.053 unabradedcondition. The HAR-A4 (in u3) G. H. Curtis,Nature 354, 145 (1991). 4980-02 1 0.0289 56.6 0.648 0.063 9. A. Deinoand R. Potts,J. Geophys. Res. 95, 8453 excavationsyielded a small numberof bi- (1990). 4980-03 1 0.0029 97.6 0.632 0.163 faces in situ as well as a large numberof 10. J. E. Kalbet al., Nature298, 25 (1982). 4980-04 1 0.0119 30.0 0.731 0.072 flakesand flakefragments, a significantpro- 11. J. S. Aigner,Anthropos (Brno) 23, 157 (1986). 4980-05 1 0.0000 75.8 0.597 0.054 2 12. M.Zhou and C. K. Ho, Geol. Soc. Am. Spec. Pap. 4980-06 1 0.0359 53.9 0.511 0.033 portionof which werequite small (1 to cm 242 (1990), p. 69. 4980-07 3 0.0040 46.3 0.389 0.073 in maximumdimension). These relations 13. P. Andrews,Cour. Forsch. Inst.Senckenberg 69, 4980-08 3 0.0105 22.8 0.565 0.075 indicatethat sortingbefore burial was min- 167 (1984). 4980-09 3 0.0403 70.1 1.681 0.033 imal. 14. R. J. Clarke,J. Hum.Evol. 19, 699 (1990). MA90-23 15. G. P. Rightmire,The Evolutionof Homo erectus The local shift from Oldowan to (CambridgeUniv. Press, New York,1990). 5005-01 3 0.0536 79.1 0.701 0.044 Acheuleanassemblages observed in the Mid- 16. B. Asfaw et al., Nature360, 732 (1992). 0.0062 0.687 0.046 5005-02 3 65.5 dle Awashcomes substantially later than the 17. M. D. Leakey, Excavations in Beds I and II, 5005-03 3 1.0680 89.3 0.871 0.080 1960-1963, vol. 3 of OlduvaiGorge (Cambridge 5005-04 3 0.0102 94.6 0.776 0.043 appearanceof the Acheuleanin Africa,dated Univ.Press, Cambridge,1971). 5005-05 5 0.2717 46.6 0.748 0.047 elsewhereminimally at 1.5 Ma (16). Old- 18. J. D. Clark,The Prehistory of Africa(Praeger, New owan and Acheuleanassemblages found in York,1970). 19. __, in After the ,G. LI. contemporaryOlduvai bed II depositswere Isaac and K. Butzer,Eds. (Mouton,The Hague, attributedfirst to differentcontemporary hom- Netherlands,1975), pp. 608-611. -0.6 Ma is based on the structural, bio- inid taxa (17) and then to differenttechno- 20. R. L. Hay, Geology of the OlduvaiGorge (Univ.of CaliforniaPress, Berkeley,1976). stratigraphic, and archaeological consider- logicalresponses to differenthabitat (18-20) 21. D. Stiles, Afr.Archaeol. Rev. 9, 1 (1991). ations outlined above. These tie the fossils or raw materialuse (21). The local Middle 22. G. LI.Isaac, Publ. Mus.Arqueol. (5th Panafrican to the adjacent unit u, whose best age is Awash.assemblage change comes within unit Congress, Santa Cruz de Tenerife,Spain) 6 (2), 0.64 + 0.04 Ma and whose minimum age is u; there is no faunaldifference between the 135 (1966). 23. , Olorgesailie:Archaeological Studies of a 0.55 ? 0.03 Ma. This placement has sig- Oldowanlevels (ul andu2) andthe overlying MiddlePleistocene LakeBasin in Kenya(Univ. of nificant implications for the interpretation unit in which Acheulean artifactsbecome Chicago Press, Chicago, 1977). of hominid evolution across the Middle pervasive(u3). Thereis, however,a progres- 24. R. Potts, J. Hum.Evol. 18, 477 (1989). 25. R. Blumenschine,ibid. 21, 451 (1991). Pleistocene. Age estimates for south and sive changein the sedimentaryregime: The 26. J. A. J. Gowlett,World Archaeol. 20, 13 (1988). east Asian representatives of Homo erectus wadi fan environmentbecame increasingly 27. G. WoldeGabrielet al., J. FieldArchaeol. 19, 471 overlap the Bodo age estimate (11, 12). If dominant,and sandsaccounted for a greater (1992). 28. MiddleAwash projectwork at Bouriin 1992 has these estimates for Asian fossils are correct, volume of the section in subunitu3. The yielded large collections of vertebrate(including evolution within African and Asian Homo MiddleAwash changefrom core (chopper)- hominid)and archaeological remains that obvi- erectus may have followed different trajec- based Oldowanassemblages to assemblages ously antedate the Bodo assemblages on bio- chronologicalgrounds, as predicted in (4). tories, with advanced features appearing with large quantitiesof Acheulean bifaces 29. C. C. Swisher IIIetal., Science 263, 1118 (1994). earlier in Africa (13-15). Alternatively, appearsto reflectthe use of differenttechnol- 30. A. Deino, L. Tauxe, M. Monaghan,R. J. Drake,J. other specimens interpreted to have ogies in differentgeographic settings. Isaac Geol. 98, 567 (1990). evolved from Homo erectus (such as noted the clusteringof Acheuleantools 31. S. D. Samson and E. C. Alexander,Chem. Geol. (22) (Isotope Geosci. Sect.) 66, 27 (1987). Arago, Petralona, and Yunxian) may be in areaswith a sandysubstratum, suggesting 32. All analyses were by direct-currentargon plasma substantially older than generally thought mechanismsof hominidpreference and hy- spectrometryon pure bulk glass separates. The

SCIENCE * VOL. 264 * 24 JUNE 1994 1909

This content downloaded on Sun, 30 Dec 2012 10:38:51 AM All use subject to JSTOR Terms and Conditions glass in samples MA90-7,MA90-20, and MA90-23 with an inverse variance weighting factor that system blank volumes of 40Ar, 39Ar, 38Ar, 37Ar, containedsignificant feldspar microlites. To insure uses deviationsabout a weighted mean to deter- and 36Arare 4, 2, 0.08, 0.3, and 0.3 x 10-17 mol, thatonly pure glass (>99%)was analyzed,a special mine the weighted error(30, 31). Individualgrain respectively. Sanidine from the Fish Canyon tuff procedurewas developed to separate even finely errors reflect errors in the fluence monitorcon- was used as the neutron fluence monitor, which disseminatedimpurities from the glass. The final stant (J) and in the determinationof Ar isotopic has a reference age of 27.84 Ma (30). The con- puritywas determinedby the inspectionof pow- ratios,which in turnpropagate errorsin discrimi- stants used in this study are as follows: (XA+ A ' dered bulkglass separates (<325 mesh; 45 ,um) nationand inAr beam intensitiesfrom sample and = 0.581 x 10-10/year; XA3= 4.962 x 1010/year; withtransmitted light microscopy and by compari- blank (30). The data in italics are excluded from 40 K per degree kelvin = 1.167 x 10-4 mol/mol). son of the DCP bulkglass Fe, Ca, Al, Mn,and Ti the mean age. Feldsparswere irradiatedfor 0.5 34. We thank the National Science Foundation, the concentrationswith those determinedby graindis- hourat 8 MWat the Omega West researchreactor National Geographic Society, the Japan Society crete electron microprobeanalyses. These steps of the Los Alamos NationalLaboratory, which has for the Promotion of Science, the University of were appliedto allsamples of this studyto facilitate a fast neutronfluence of 5.7 x 1013 neutronsper California Collaborative Research Program of the comparisonsand correlations.Concentrations of Fe, centimeterper second. Cadmiumshielding was Institute of Geophysics and Planetary Physics at Ca, Al, Mn,and Tiwere determinedby comparison used to reduce the thermalneutron production of Los Alamos National Laboratory, the Ethiopian of unknownintensities to calibrationcurves on the 40Ar.After irradiation,the samples were trans- Ministryof Culture and Sports Affairs, the National basis of six internationallyrecognized rock stan- ferred to a copper holder and loaded onto the Museum of Ethiopia, the Afar people, A. Ade- dards.All other concentrations were determinedby extractionline for overnightbakeout at -200?C. massu, B. Latimer, S. Yirga, G. Hundie, A. Ne- the methodof standardsaddition with the use of a Samples were fused with an 8-W argon-ionlaser gash, Z. Assefa, and C. Tilahun for making the multi-elementspike solutiondesigned specifically (Mass Analyser Products, Nottingham, United 1990 Middle Awash research possible. We also forsilicic compositions. Kingdom).Abundances of the argon isotopes thank F. C. Howell and B. Dalrymple for helpful 33. The Ca/K ratiois derived by the multiplicationof were measured on a MAP-215noble gas spec- suggestions on the paper. the measured37Ar/39Ar ratio by 1.96. The weight- trometerfitted with a Balzers electron multiplier ed mean age and error(a; 1 sigma) are computed operating at a gain of about 20,000. Typical 8 February 1994; accepted 2 May 1994

MCM-22:A MolecularSieve with Two Independent calcined in air at 540?Cfor 16 hours. This calcined product, HO.033NaO043 (AlO005- MultidimensionalChannel Systems Bo.071sio.924)02, is referredto as MCM-22. Electron diffraction (ED) photographs Michael E. Leonowicz, Jeffrey A. Lawton,*Stephen L. Lawton,t indicatedthat the unit cell had hexagonal Mae K. Rubin symmetry with no systematic absences. They exhibited Laue symmetryconsistent The molecular sieve MCM-22 contains structural features previously unobserved in this with the eight hexagonal space groups class of materials. Its framework topology, derived from high-resolution electron micro- P6/mmm, P62m, P6m2, P6mm, P622, graphs and refined with synchrotron x-ray diffraction powder data, contains two indepen- P3lm, P3lm, and P312. The XRD data for dent pore systems, both of which are accessed through rings composed of ten tetrahedral the structurerefinement of the calcined (T) atoms (such as Si, Al, and B). One of these pore systems is defined by two-dimensional, samplewere collected on beamlineX7A at sinusoidal channels. The other consists of large supercages whose inner free diameter, the National SynchrotronLight Source in 7.1 angstroms, is defined by 12 T-O species (12-rings).and whose inner height is 18.2 Brookhaven National Laboratory (3). angstroms. These coexisting pore systems may provide opportunities for a wide variety of These data were successfullyindexed on a catalytic applications in the petrochemical and refining industries. Another structural fea- hexagonallattice and were fully consistent ture is an unusual -T-O-T- chain that passes through the center of a modified dodecasil-1 H with a spacegroup that exhibitsno system- [435663] cage. atic extinctions. The refinedunit cell pa- rameterswere a = 14.1145(8) A and c = 24.8822(18)A (4). The structurewas determinedwith high- Molecularsieves are crystalline, microporous very thin sheets and that boron MCM-22 resolutionelectron micrograph (HREM) lat- materialswith frameworksof tetrahedralat- crystallitesare typicallythicker than those tice imagesand modelbuilding. Trial models oms that are interconnectedin threedimen- fromaluminum syntheses. Boron-containing wererefined by the programDLS (5), andthe sionsthrough oxygen atoms. Their pore struc- sampleswere chosen for structuralstudies model whose calculatedXRD pattem most turemakes them usefulin the petrochemical becausetheir large size gave rise to powder closely resembledthe experimentalpattem and refiningindustries for catalysisand sepa- x-raydiffraction (XRD) patterns with sharper was subjectedto a Rietveldrefinement (6) rationprocesses. Synthesis is usuallycarried peaks. (Fig.1). Finalatomic coordinates are given in out underhydrothermal conditions. A direct- The MCM-22sample used in this study Table 1. Thereare 72 T atomsper unit cell. ing agent, if presentin the reactiongel, may was hydrothermallysynthesized by proce- As seen in projectionsof the framework be usedto influencethe formationand geom- dures describedelsewhere (1) from Hi-Sil topologyfor MCM-22 (Figs. 2 and 3), the etryof the intemal pore system.The molec- precipitatedsilica (87% SiO2), hexameth- [001]projection is remarkablysimilar to that ular sieve MCM-22 was synthesizedin this yleneimine (HMI), triethanolamine,boric of the silica clathratedodecasil-1H (DOH) manner,with hexahydro-1H-azepine(hexa- acid (H3BO3),and sodiumhydroxide (50% (7) (Fig. 3). Topologically,the two frame- methyleneimine)as a directingagent. We NaOH solution). Triethanolamine(320 g) work connectivitiesare related.The DOH propose a model for the structureof this wasadded to a solutioncontaining 14.0 g of structurecontains hexagonal sheets of [435663] molecularsieve. H3BO3, 6.3 g of NaOH, and 160g of H2O; cages (8) (Fig. 4A) that share 4-ring faces MCM-22may be synthesizedas an alumi- 96.0 g of Hi-Sil were then added, followed with adjacentcages within the sheet. Large nosilicateor borosilicate.Scanning electron by 42.0 g of HMI. Aftervigorous stirring for 12-ringcavities bordered by six joined cages micrographsrevealed that it crystallizesas 1 hour, 100 g of this slurrywere transferred arecreated within these sheets. These sheets, to a Teflon-linedstainless steel autoclave. in tum, join along the unit cell c direction MobilResearch and DevelopmentCorporation, P.O. The vessel was placed in an oil bath, and throughsingle 6-ringsthat cap the 12-ring Box 480, Paulsboro,NJ 08066, USA. the reactionmixture was crystallizedunder cavitiesat eitherend, givingthe cavity[68512] *Present address: Program in Cell and Molecular staticconditions at 120?Cfor 230 days.The topology.Adjacent sheets are also directly Biology,Baylor College of Medicine,One BaylorPla- za, Houston,TX 77030, USA. solid product was washed with water and connectedthrough bonds between apicalT tTo whom correspondence should be addressed. dried overnight at 120?C (2) and then atomson the threefoldaxes of [435663]cages. 1910 SCIENCE * VOL. 264 * 24 JUNE1994

This content downloaded on Sun, 30 Dec 2012 10:38:51 AM All use subject to JSTOR Terms and Conditions