Alan L. DeinoBerkeley GeochronologyCenter, 2455 Ridge Road,Berkeley, California 94709,U.S.A. E-mail: al@bgc.org

Andrew HillDepartment of Anthropology,Yale University, Box 208277,New Haven, Connecticut06520, U.S.A. E-mail:andrew.hill@yale.edu

Received11 September 2001and accepted17 September 2001

Keywords: 40Ar/39Ar dating,Chemeron Formation,BPRP#2, Homo, Pliocene,Tugen Hills, Kenya.

40Ar/39Ar dating of Chemeron Formationstrata encompassing the site of hominidKNM-BC 1, Tugen Hills, Kenya

A fossil hominid temporal bone (KNM-BC 1) from surfaceexposures at Baringo Paleontological Research Project site BPRP#2in the Chemeron Formation outcropping in a tributary drainageof the Kapthurin River west of Lake Baringo, Kenya has beenattributed to Homo sp. indet. K-feldspar phenocrysts from lapillituffs bracketing the inferred fossiliferous horizon yield single-crystal40Ar/39Ar ages of 2·456�0·006 and 2·393�0·013 Ma. These agedeterminations are supported by stratigraphically consistent ages onhigher tuff horizons and from nearby sections. In addition, new40Ar/39Ar ages on tuffaceous units near the base and top of theformation along the Kapthurin River yield 3·19�0·03 and1·60�0·05 Ma respectively. The base of the formation along theKapthurin River is thus �0·5 Ma younger than the uppermostChemeron Formation strata exposed at Tabarin, 23 km to thenorth-northwest. The upper half of the formation along theKapthurin River was deposited at an average rate of �11 cm/ka,compared to 21–23 cm/ka at Tabarin.

� 2002 Academic Press

Journal of Human Evolution (2002) 42, 141–151doi:10.1006/jhev.2001.0522Available online at http://www.idealibrary.com on

Introduction

A fossil hominid temporal bone (KNM-BC1) from Baringo Paleontological ResearchProject site BPRP#2 (EAGRU site JM85)in the Chemeron Formation southeast ofLake Baringo, Kenya originally describedas family Hominidae gen. et sp. indet.(Martyn, 1967; Tobias, 1967), received themore definitive taxonomic assignment ofHomo sp. indet. (Hill et al., 1992; Sherwoodet al., 1996, 2002). In Hill et al. (1992) weassigned an age of 2·43�0·02 Ma to a tuff

lying several meters below the probablefossiliferous horizon. Additional 40Ar/39Arradiometric ages are reported here from thesite area, which provide a more com-prehensive record of the age of thesesediments. We also report on age determi-nations from tuffs near the base and top ofthe Chemeron Formation in the KapthurinRiver valley.

0047–2484/02/010141+11$35.00/0

Geologic setting

The Tugen Hills, a structural horst withinthe Kenya Rift west of Lake Baringo, Kenya,expose on their eastern foothills discontinu-ous outcrops of a sequence of terrigineousand lacustrine sediments and tuffs of Plio–Pleistocene age. These beds originally con-stituted a portion of Gregory’s Kamasiasediments (Gregory, 1896, 1921), subse-quently studied by Fuchs (1934, 1950).Much later McCall et al. (1967) recognizedthat the Kamasia sediments were two dis-tinct lithological formations separated byan angular unconformity and, as theythought, a lava. The lower unit was infor-mally called the Chemeron beds; the upper,the Kapthurin beds. The Chemeron bedsreferred to the sequence of sediments lyingon a basalt, now termed the KaparainaBasalts Formation, and exposed by a seriesof rivers, such as the Kapthurin River and

� 2002 Academic Press

142 . . .

N

0 0.5 km

Site K002 Section �A�

KENYA

Baringo

Section �B�

Kapthurin River

Figure 1. Location of fossil locality BPRP#2, 40Ar/39Ardating sample localities, and section lines.

Chemeron River itself, that drain from theTugen Hills eastward into Lake Baringo.Martyn (1967, 1969) formalized thesenames and added to the ChemeronFormation exposures an apparently separatebasin centered on Kipcherere about 10 kmfarther west. These also lay on KaparainaBasalts. Pickford (1975; see also Pickfordet al., 1983) also attributed rocks farthernorth to the Chemeron Formation, some ofwhich had originally been placed in theKaperyon Formation by Chapman (1971).

In the Kapthurin River basin, the top ofthe Chemeron Formation occurs at anangular unconformity with the overlyingKapthurin Formation. A few meters of sedi-ment and tuff are found in the ChemeronFormation immediately below this contact,which are in turn underlain by a prominentstratigraphic marker unit in the area, theNdau Trachymugearite mafic flow (Hillet al., 1985, 1986). Previous K–Ar dates onthis and another similar lava gave 1·57 and2·13 Ma (Chapman & Brook, 1978; Hillet al., 1985). The Chemeron Formation inthe Kapthurin River area disconformablyoverlies basaltic lavas and minor inter-calated sediments of the Kaparaina BasaltsFormation. K–Ar and 40Ar/39Ar ages forthis formation in the Tugen Hills typicallyrange from 5·7–5·1 Ma (Chapman &Brook, 1978; Hill et al., 1985; Deino et al.,2002).

BPRP#2 is located within an eastward-dipping structural block containing stratafrom the middle of the Chemeron Forma-tion sequence as exposed in the KapthurinRiver drainage. The sequence is dominatedby fine-grain terrigineous and lacustrinesediments, primarily mudstone and silt-stone, with diatomite, sandstone, and con-glomerate intercalations. In addition,tuffaceous material is found as reworkedshard and pumice components in sands, asdispersed lapilli and ash in diatomite ormudstone, or as discreet beds of lapilli orvitric tuff.

A distinctive, resistant 4-m-thick beddedlapilli and block tuff, informally called the‘‘Lower Tuffs’’ by Martyn (1967), supportsa prominent dip-slope throughout theBPRP#2 area. The tuff was deposited ina lacustrine environment based on theoccurrence of ripple marks and laminated,horizontal bedding within the tuff, and itsstratigraphic association: the basal portiongrades downward into diatomite, and theupper portion is overlain by fine-grainedsiltstone containing fish bones. The tuff

sequence is found throughout the KapthurinRiver drainage and beyond with similar sedi-mentological attributes and thickness, sug-gesting deposition in a relatively broad lakebasin. Deposition of the tuff probablyoccurred rapidly as the result of a singleeruptive sequence. We include in the upper-most part of the tuff 0·5 m of tuffaceous siltscontaining lapilli and blocks of pumice simi-lar to those of the underlying laminated tuff;the admixture of silt apparently recordswaning of this eruption, or reworking. Anage for the upper part of this tuff of2·43�0·02 Ma by the single-crystal 40Ar/39Ar dating method was reported in Hillet al. (1992). The location of these sectionsand of the 40Ar/39Ar sample localities aregiven in Figure 1. Stratigraphic columns areshown in Figure 2 for two measured sectionsnear BPRP#2.

143

Section �A� Section �B�

40

35

30

25

20

15

10

5

0

0

10

15

20

25

30

KAP-22.367 ±0.016 Ma

KAP-62.410 ±0.022 Ma

ALD86-1,ALD86-1BRD-6

2.456 ±0.006 Ma

KNM-BC1?

AH90-12.353 ±0.006 Ma

�Low

er T

uff

s�

KAP-32.393 ± 0.013 Ma

5

Figure 2. Stratigraphic sections measured in theChemeron Formation in the vicinity of site BPRP#2.Thickness given in meters.

Section ‘‘A’’ was measured through siteBPRP#2, while section ‘‘B’’ is locatedabout 200 m south at the confluence of thetributary leading to BPRP#2 and the mainKapthurin River channel. These sectionsoccur along strike and dip 21–27� eastward.Correlation between sections is based onlithologic comparison and structural place-ment. Tracing of strata from Section A tonear the site of Section B indicates that thissequence lies stratigraphically above theLower Tuffs. Several distinctive coarse clas-tic units, tuff beds, and a distinctive redclaystone can be recognized in both sections.

KNM-BC 1 was found on the surfaceabove the ‘‘Lower Tuffs,’’ on the dip-slopewest of the tributary to the Kapthurin Riveradjacent to the BPRP#2 site (Martyn,1967). For a full account of the circum-stances of its discovery and recognition seeHill (1999). Based on the lithology of the‘‘lithic grit’’ adhering to the fossil, Martyn(1967) traced its probable origin to a sandyhorizon upslope from the point of discovery,stratigraphically about 8� (2·4 m) above theLower Tuffs. Subsequent workers endorsedthis conclusion, and there remains nofurther doubt about the provenance of thisfossil (Hill, 1999). In the area of Section‘‘A,’’ this unit is probably represented by ahalf-meter-thick coarse sand 2·5–3·0 mabove the top of the Lower Tuffs. Datingsample KAP-3 was obtained on the eastslope of the gully directly across fromBPRP#2; it lies stratigraphically above theinferred fossiliferous horizon by about 4 m.By projection of bedding attitudes, this tuf-faceous horizon also lies stratigraphicallyabove the youngest Chemeron Formationsediments in the BPRP#2 site area on thewest slope of the gully, and thus provides afirm upper age limit for the fossil. Anotherlapilli tuff 13·1 m above KAP-3 in this samesection was also dated (AH90-1). As afinal check on the chronostratigraphy, twolapilli tuffs were dated from section B incorrelative sediments (KAP-2, KAP-6).

144 . . .

Table 1a 40Ar/39Ar analytical data, Chemeron Formation, Middle Kapthurin River, Baringo Basin,Kenya

Lab ID# Ca/K 36Ar/39Ar 40Ar*/39Ar

40Ar(mol�1014) %40Ar*

Age (Ma)�1�

Sample AH90-1: Lapilli tuff (pumice to 2 cm), with detrital matrix of sub-angular class of volcanic rock and lesserfeldspar. ‘‘A,’’ ‘‘B,’’ and ‘‘C’’ are individual lapilli.

Sample AH90-1/A:6551-01 0·013 0·00005 0·824 1·1 97·9 2·40�0·026551-02 0·044 0·00004 0·816 1·5 98·2 2·38�0·026551-03 0·019 0·00004 0·815 0·6 98·2 2·37�0·036551-04 0·043 0·00006 0·811 0·7 97·4 2·36�0·026551-05 0·033 0·00006 0·815 0·7 97·4 2·37�0·026551-06 0·065 0·00008 0·809 0·6 96·6 2·36�0·026551-07 0·046 0·00095 0·809 0·6 96·6 2·36�0·026551-08 0·012 0·00016 0·795 0·4 94·1 2·31�0·026551-09 0·058 0·00008 0·803 0·9 97·0 2·34�0·026558-01 0·037 0·00057 0·808 0·4 82·5 2·36�0·036558-02 0·046 0·00022 0·809 0·3 92·3 2·36�0·036558-03 0·040 0·00016 0·809 0·2 94·0 2·36�0·036558-04 0·047 0·00010 0·819 0·5 96·1 2·39�0·036558-06 0·022 0·00004 0·800 1·0 98·3 2·34�0·026558-07 0·013 0·00007 0·806 1·5 97·2 2·35�0·026558-08 0·012 0·00005 0·821 1·4 97·8 2·40�0·026558-09 0·037 0·00006 0·810 0·9 97·5 2·36�0·026558-10 0·012 0·00009 0·812 0·7 96·4 2·37�0·026558-11 0·045 0·00023 0·820 1·8 91·9 2·39�0·026558-12 0·012 0·00006 0·811 0·8 97·3 2·37�0·02

Wtd. Ave.= 2·361�0·009Omitted (<80% 40Ar*):6558-05 0·045 0·00109 0·798 2·7 71·2 2·32�0·02

Sample AH90-1/B6552-01 0·019 0·00004 0·811 1·2 98·0 2·37�0·026552-02 0·019 0·00007 0·803 1·1 96·9 2·35�0·026552-03 0·015 0·00004 0·805 1·4 98·1 2·35�0·026552-04 0·066 0·00039 0·809 1·4 87·3 2·37�0·026552-05 0·070 0·00048 0·801 2·3 84·7 2·34�0·026552-06 0·068 0·00041 0·805 0·7 86·8 2·35�0·036553-01 0·065 0·00054 0·795 1·2 83·2 2·32�0·026553-03 0·013 0·00006 0·812 0·4 97·3 2·38�0·036553-04 0·060 0·00006 0·803 1·0 97·6 2·35�0·02

Wtd Ave.= 2·347�0·009Omitted (<80% 40Ar*):6553-02 0·061 0·00092 0·798 1·1 74·5 2·34�0·02

Sample AH90-1/C6555-01 0·016 0·00007 0·804 1·0 96·9 2·35�0·026555-02 0·016 0·00003 0·804 2·2 98·5 2·35�0·026555-03 0·013 0·00003 0·818 1·3 98·4 2·39�0·026555-04 0·016 0·00007 0·800 0·7 97·2 2·34�0·02

Wtd. Ave.= 2·351�0·012Overall weighted mean of samples AH90-1/A, B, and C= 2·353�0·006

Sample KAP-2; Lapilli (<1 cm) from laminated thin white tuff and sandstone interbeds 2–15 cm total thickness.1444-01 0·0007 0·00013 0·968 1·5 96·0 2·36�0·021444-02 0·0030 0·00004 0·964 1·7 98·6 2·35�0·011444-03 0·0063 0·00041 0·982 1·2 89·0 2·40�0·021444-04 0·0025 0·00007 0·959 2·0 97·7 2·34�0·01

145

Table 1a Continued

Lab ID# Ca/K 36Ar/39Ar 40Ar*/39Ar

40Ar(mol�1014) %40Ar*

Age (Ma)�1�

Sample KAP-2; Lapilli (<1 cm) from laminated thin white tuff and sandstone interbeds 2–15 cm total thickness.1444-05 0·0019 0·00002 0·971 1·2 99·2 2·37�0·021444-06 0·0111 0·00037 0·988 1·4 90·0 2·41�0·021444-08 0·0004 0·00009 0·986 1·2 97·4 2·41�0·02

Wtd. Ave.= 2·367�0·016Omitted (<80% 40Ar*):1444-07 0·0006 0·00207 0·986 2·6 61·7 2·41�0·02

Sample KAP-3: Fragile sparse lapilli up to 4 cm in diameter dispersed in a 5 cm red mudstone bed, directlyopposite JM85 location.1447-01 0·036 0·00006 0·984 4·0 98·2 2·40�0·011447-02 0·023 0·00006 0·973 2·3 98·3 2·38�0·011447-03 0·031 0·00008 0·980 4·0 97·8 2·39�0·011447-04 0·039 0·00020 0·994 3·0 94·5 2·43�0·011448-01 0·038 0·00008 0·978 1·7 97·7 2·39�0·011448-02 0·029 0·00001 0·991 2·1 99·6 2·42�0·011448-03 0·031 0·00005 0·989 2·3 98·6 2·41�0·011448-04 0·021 0·00005 0·981 1·6 98·7 2·40�0·011448-05 0·039 0·00009 0·978 2·7 97·4 2·39�0·01

Wtd. Ave.= 2·393�0·013Omitted (Greater than two standard deviations beyond mean):1447-05 0·037 0·00004 1·023 2·5 99·0 2·50�0·01

Sample KAP-6: 15-cm pumice block from top of a white diatomaceous siltstone with sparse dispersed lapili.4701C-01 0·021 0·00012 0·630 1·1 91·8 2·43�0·034701C-02 0·031 0·00022 0·616 0·6 87·9 2·37�0·054701C-05 0·003 0·00036 0·639 2·2 83·6 2·46�0·024701C-06 0·040 0·00017 0·626 1·6 90·2 2·41�0·024701C-07 0·013 0·00022 0·616 1·7 87·7 2·37�0·02

Wtd. Ave.= 2·410�0·022Omitted (<80% 40Ar*):4701C-03 0·035 0·00171 0·626 2·9 54·4 2·40�0·034701C-04 0·062 0·00059 0·630 1·2 76·5 2·42�0·034701C-08 0·154 0·00707 0·568 0·7 21·3 2·18�0·20

Sample ALD86-1B: Pumice block from near top of the ‘‘Lower Tuffs.’’7387/2-01 0·053 0·00007 1·008 5·1 98·1 2·45�0·017387/2-02 0·051 0·00004 1·008 8·4 99·0 2·45�0·017387/2-03 0·057 0·00007 1·003 6·8 98·3 2·44�0·017387/2-04 0·020 0·00010 1·003 3·2 97·2 2·44�0·017389/2-01 0·050 0·00013 0·997 3·9 96·4 2·44�0·017389/2-02 0·021 0·00005 1·004 4·0 98·5 2·45�0·017389/2-03 0·055 0·00005 1·004 6·7 98·7 2·45�0·017389/2-04 0·021 0·00014 1·003 7·2 96·2 2·45�0·017389/2-05 0·049 0·00023 1·002 5·9 93·8 2·45�0·01

Wtd. Ave.= 2·451�0·008

Sample ALD86-5: Vitric fine-grained tuff approximately 0·6 m thick at top of Chemeron Formation, just aboveNdau Trachymugearite along the south bank of the Kapthurin River.292-02 0·112 0·00007 0·329 0·9 95·4 1·65�0·08292-03 0·050 0·00009 0·307 0·8 92·3 1·54�0·08292-04 0·082 0·00008 0·322 0·8 94·1 1·61�0·08

Wtd. Ave.= 1·60�0·05Omitted (<80% 40Ar*):292-01 0·107 0·00065 0·324 1·2 63·4 1·62�0·10

146 . . .

Table 1a Continued

Lab ID# Ca/K 36Ar/39Ar 40Ar*/39Ar

40Ar(mol�1014) %40Ar*

Age (Ma)�1�

Sample AD/RD-3: Vitric fine-grained tuff 0·1 m thick, with 5–10% lapilli to 2 cm, 19·3 m above base of ChemeronFormation where it overlies Kaparaina Basalt along the south bank of the Kapthurin R.279-01 0·050 0·00006 0·652 1·9 97·9 3·21�0·05279-02 0·051 0·00014 0·638 3·1 94·4 3·14�0·04279-03 0·053 0·00002 0·647 1·8 99·3 3·18�0·06279-04 0·081 0·00015 0·641 1·7 93·9 3·16�0·07279-06 0·022 0·00016 0·647 1·0 93·2 3·18�0·05279-07 0·040 0·00011 0·657 1·6 95·6 3·23�0·04

Wtd. Ave.= 3·19�0·03Omitted (<80% 40Ar*):279-05 0·019 0·00935 0·510 4·2 15·6 2·51�0·22

Notes: Errors in age quoted for individual runs are 1� analytical uncertainty. Weighted averages are calculatedusing the inverse variance as the weighting factor (Taylor, 1982), while errors in the weighted averges are 1�standard error of the mean (Samson and Alexander, 1987) and incorporate error in J, the neutron flux irradiationparameter. Ca/K is calculated from 37Ar/39Ar using a multiplier of 1·96. 40Ar* refers to radiogenic argon. ‘‘Moles40Ar’’ refers to the estimated total moles of 40Ar released during fusion based on spectrometer sensitivityconsiderations. �=5·543�10�10 y�1.

Table 1b 40Ar/39Ar irradiation facilities and parameters

LabID# Reactor

Irrad. time(hr)

MonitorMineral

J�10�3

(36Ar/37Ar)Ca

�10�4(39Ar/37Ar)Ca

�10�4(40Ar/39Ar)K

�10�2

279 B 10 MMhb-1 2·726�0·02 2·59�0·06 9·0�0·3 0·86�0·07292 B 10 MMhb-1 2·772�0·02 2·59�0·06 9·0�0·3 0·86�0·07

1444 L 2 FCT 1·352�0·007 2·58�0·06 6·7�0·3 0·08�0·071447 L 2 FCT 1·352�0·007 2·58�0·06 6·7�0·3 0·08�0·071448 L 2 FCT 1·352�0·007 2·58�0·06 6·7�0·3 0·08�0·074701 L* 3 PF 2·128�0·010 2·58�0·06 6·7�0·3 2·08�0·136551 D 7 FCT 1·612�0·005 2·64�0·02 6·7�0·4 0·45�0·506552 D 7 FCT 1·618�0·005 2·64�0·02 6·7�0·4 0·45�0·506553 D 7 FCT 1·623�0·005 2·64�0·02 6·7�0·4 0·45�0·506555 D 7 FCT 1·619�0·005 2·64�0·02 6·7�0·4 0·45�0·506558 D 7 FCT 1·615�0·005 2·64�0·02 6·7�0·4 0·45�0·507387 O 14 FCT 1·358�0·004 2·64�0·02 6·7�0·4 0·07�0·03

Reactors: ‘‘B’’=Central thimble facility of the UCB TRIGA reactor in Berkeley, California, operated at 1 MW;‘‘L’’=Hydraulic rabbit facility of the Los Alamos Omega West reactor operated at 5 MW; ‘‘D’’=Central thimblefacility of the USGS TRIGA reactor in Denver, Colorado, operated at 1 MW; ‘‘O’’=CLICIT Dummy fuelelement position in the Oregon State University TRIGA reactor, operated at 360 KW. *Indicates no Cd shielding;all others are Cd shielded. J=neutron fluence parameter (36Ar/37Ar)Ca (39Ar/37Ar)Ca are isotopic ratios derivedfrom irradiation of Ca in the sample (40Ar/39Ar)K is an isotopic ratio derived from irradiation of K in the sample.Monitor minerals: MMhb-1=McClure Mountain Complex hornblende, reference age 523·1 Ma (Renne et al.,1988); FCT=Fish Canyon Tuff sanidine, reference age 28·02 Ma (Renne et al., 1988); PF=sanidine fromash-flow tuff of the Pahranagat Formation, reference age 22·785�0·009 determined against FCT at 28·02 Ma(Best et al., 1995, modified for Renne et al., 1998).

40Ar/39Ar datingK-feldspar phenocrysts extracted from fivepumice tuff samples were analyzed by the

single-crystal, laser-fusion, 40Ar/39Ar datingmethod. Pumice were gently crushed andsieved between the 12 to 28 mesh size

147

2.262.28

2.302.32

2.342.36

2.382.40

2.422.44

050

100150200250300350

500450400

550600650

Rel

ativ

e pr

obab

ilit

y

Age (Ma)

SampleAH90-1

Overall

CB

A

Figure 3. Age-probability spectra for lapilli A, B, and Cfrom sample AH90-1. The vertical axis of the plot is arelative probability measure of obtaining a particularage for given sample, based on the sum of the gaussianerrors of the individual single-crystal analyses. Thelocation of the weighted mean and one-sigma standarderror of the weighted mean are also shown for eachlapillus, and for the overall distribution.

fraction. Feldspar crystals were removedand treated with dilute HCl, HF, and dis-tilled water in an ultrasonic bath to removeadhered matrix, and then hand-selected toobtain pristine, inclusion-free material.Irradiation facilities and parameters arelisted in Table 1B. After a sufficient intervalto permit radiological cooling to acceptablelevels following irradiation, samples wereloaded in a ultra-high-vacuum samplechamber and baked at 300�C overnight.Grains were first illuminated with a low-power, diffuse laser beam (Ar-ion orNd-YAG) to desorb surficial atmosphericargon, which was immediately pumpedaway. The feldspars were then fused with afocused laser beam at high power (8–35Watts). Gases released by this step werescrubbed of reactive chemical species byexposure to SAES getters for 3–5 min, thenanalyzed for five Ar isotopes on a MAP 215low-blank mass spectrometer. Furtherdetails of analytical procedures are providedin Deino & Potts (1990).

Results of these analyses are presented inTable 1. Radiogenic 40Ar content (40Ar*)ranged from 15·6 to 99·6%, with the greatmajority of grains yielding >90% 40Ar*. Therelatively high atmospheric yields of somegrains could arise from a number of factors,including alteration, fluid or glass inclu-sions, incomplete fusion, and relativelysmall grain size. Those few analyses with40Ar* less than an arbitrary 80% cut-off

were culled from the data set to mitigate theinfluence of these factors. Elimination ofthese analyses affects the weighted meanages by less than 1%, as runs with lowproportions of radiogenic argon tend tohave higher uncertainties due to the largeatmospheric corrections required, and so aretypically not influential in a weighted meancalculation.

One other analysis (1447-05) fromsample KAP-3 was omitted because of ananomalously old age that fell two stan-dard deviations beyond the mean. This may

represent an analytical artefact or a geologiccontaminant. Elimination of this runchanged the weighted mean age for thissample by less than 0·5%. As a check onanalytical reproducibility and age variabilitybetween pumice lapilli, anorthoclase fromthree lapilli from the same tuff at a singlelocality (AH90-1) were extracted andanalyzed separately (Table 1). The agepopulation systematics are illustrated inFigure 3 using age-probability diagrams.

These diagrams are essentially the sum ofthe gaussian errors of the individual single-crystal analysis (Deino & Potts, 1992). Theyare similar to conventional histograms thatshow the frequency of occurrence of a speci-fied age against time, with the differencethat the age-probability spectrum specifi-cally accommodates the variable analyticaluncertainties inherent in argon-isotopic agedata sets.

The age-probability diagram shows thatlapilli A and B have unimodal age distribu-tions with n of 20 and 9, respectively. C,however, is bimodal, though this is probablydue to insufficient replicate analyses since

148 . . .

only four anorthoclase grains were availablefor dating. Although there is some indica-tion in A and possibly C of skewing towardolder ages, the effect on the weighted meanage is negligible since the median and mode,fairly robust measures of central tendencywith small or skewed populations, both fallwithin a few tenths of a percent of theweighted-mean age (Table 2). The maxi-mum difference between lapilli mean ages is0·01 Ma, or 0·4%. This is within statederrors, and none of the lapilli mean agescould be differentiated using a Student’st-test at the 95% confidence level. The uni-modal nature of populations A and B: theagreement in age of all samples; concord-ance of the mean, median, and mode; andreproducibility on the order of a few tenthsof a percent attest to the general analyticalreliability of the dating results.

As a check on the age we obtained pre-viously for the ‘‘Lower Tuffs’’ (Hill et al.,1992), we re-separated and analyzedphenocrysts from one of the originalsamples, pumice block ALD86-1 (hereidentified as ALD86-1B). The overallweighted mean age reported by Hill et al.(1992) for the ‘‘Lower Tuffs’’ was2·45�0·01 Ma, or 2·43�0·02 Ma by theisochron approach (ages are adjusted for arevised age of the standard, see Renne et al.,1988). In the interim since these earlyanalyses were performed, much of theanalytical methodology has changed; most

Population statistics for separate age analyses of three lapilli from sampleAH90-1

Lapillus nWeighted mean

age (Ma)�1� S.E.M.

(Ma)*Median

agePrincipal

mode

A 20 2·361 0·009 2·357 2·357B 10 2·347 0·009 2·349 2·349C 4 2·351 0·012 2·348 2·348

Weighted mean of A, B, and C:34 2·353 0·006 2·356 2·349

*Standard error of the mean, incorporating uncertainty in J, the neutron fluxparameter.

Table 2

importantly, the new samples were irradi-ated in a different reactor, calibrated withdifferent monitor minerals, and measuredon a different mass spectrometer and extrac-tion line, using considerably modified gas-extraction and data-reduction techniques.Nevertheless, the new analyses yielded avery consistent age population with aweighted mean age of 2·451�0·008 Ma,within analytical error of the earlier result.The final weighted mean age of the priorand current results is 2·456�0·006 Ma(uncertainty is 1� standard error of theweighted mean; Taylor, 1982).

Lapilli tuff sample KAP-3 provides ayounger age limit for the stratigraphic inter-val from which the fossil must havebeen derived, yielding 2·393�0·013 Ma(Table 1 and Figure 2). A t-test indicatesthat this age is significantly different fromthat of the marker tuff (ALD86-1B) at the95% confidence level. The mean sedimenta-tion rate between these tuffs is 12�2 cm/ka(uncertainty calculated as the square root ofthe sum of the squares of the errors of thedating results), based on a measured strati-graphic thickness of 8·0 m over the 63 kyrtime interval (Figure 4). This interval con-sists dominantly of lacustrine mudstonesand siltstones and lacks any sign of a hiatus.

A lapilli tuff (sample AH90-1) 12·9 mabove KAP-3 in section ‘‘A’’ yielded astratigraphically consistent age of 2·353�0·006 Ma. An overall sedimentation rate

149

2.342.35

2.362.37

2.382.39

2.402.41

2.422.48

2.472.46

2.452.44

2.43�14

�12

�10

�8

�6

�4

�2

4

2

0

6

8

10

Dis

tan

ce f

rom

ref

eren

ce h

oriz

on (

m)

Age (Ma)

Interpolated fossil age

Probable KNM-BC1 Horizon

AH90-1

KAP-2

KAP-3KAP-6

ALD86-1,ALD86-1B,and RD-6

12 ± 2 cm/ka

12 ± 11cm/ka

32 ± 15 cm/ka

Figure 4. 40Ar/39Ar ages plotted as a function ofstratigraphic height, showing mean sedimentation rates.The base of a correlative conglomerate sequence foundin section ‘‘A’’ at 20·7 m and in section ‘‘B’’ at 10·0 mwas chosen as a stratigraphic reference horizon forcomparing the two sections on this diagram.

rocks conformably overlying the NdauTrachymugerite. This thin sequence is dis-conformably overlain by the KapthurinFormation. The approximately 90-m-thicksection between the hominid temporal leveland the top of the formation was depositedin 0·8 Ma, indicating a mean sedimentationrate of �11 cm/ka.

The base of the Chemeron Formationalong the Kapthurin River must be olderthan 3·19�0·03 Ma, as constrained by a40Ar/39Ar age of pumice from a fine-grainedtuff occurring in a section of brownmudstones 19 m above the contact withweathered basalt of the Kaparaina BasaltsFormation (sample AD/RD-3). A sedimen-tation rate for the lower part of the forma-tion along the Kapthurin River cannot beestimated due to structural complicationsand incomplete exposure. However, byanalogy to the sedimentation rate deter-mined for the upper part of the section, theundated 19 m interval may be roughly0·15 kyr in duration. Thus ChemeronFormation deposition in this region mayhave begun at about 3·3 Ma.

calculated for the interval between thesetuff horizons is 32�15 cm/ka, more thantwice that obtained for the lower part of thesection, but within the large uncertainties ofthese determinations.

Two tuffs dated at meter level 5 and 10 insection ‘‘B’’ yielded ages of 2·410�0·022(KAP-6) and 2·367�0·016 Ma (KAP-2),stratigraphically consistent with the agedeterminations from section ‘‘A’’ if thelithostratigraphic correlations are correct(Figure 2). A sedimentation rate calculatedfor the stratigraphic interval between thesetuffs is 12�11 cm/ka, equal in magnitudeto the rate calculated for the lower part ofsection ‘‘A.’’

In addition to our analysis of tuffs near thehominid temporal level, we have obtainedradiometric age constraints for the top andbottom of the Chemeron Formation in theKapthurin River valley. The top of theformation is 1·60�0·05 Ma, based on asingle-crystal 40Ar/39Ar age on a fine-grainedtuff at the top of 2 m of sedimentary

Discussion

KNM-BC 1 was found on a dip slope com-prising the western wall of a small gully cutby a tributary of the Kapthurin River.The dip slope is underlain by a prominenttuffaceous marker unit with an age of2·456�0·006 Ma. A tuff on the oppositeside of the gully stratigraphically above theprojection of the dip slope yielded an age of2·393�0·013 Ma. These ages are statisti-cally distinguishable at the 95% confidencelevel, and constrain the fossil to a 63 kainterval. Further, if attribution of the fossilto a sandstone bed about 2·4 m above themarker tuff is correct, then its interpolatedage is about 2·42 Ma (Figure 4). Study ofnearby sections in the same stratigraphicinterval provides additional verification ofthis general age assignment.

150 . . .

Acknowledgements

This research is part of the work of theBaringo Paleontological Research Project(BPRP), based at Yale University, and car-ried out in collaboration with the NationalMuseums of Kenya. We thank theGovernment of the Republic of Kenya forpermission to carry out research in Kenya(Permit OP/13/001/C 1391/issued to AH),and permission to excavate from theMinister for Home Affairs and NationalHeritage. BPRP has been supported bygrants to AH from NSF (most recentlySBR-9208903), the Louise H. and David S.Ingalls Foundation, the Louise BrownFoundation, Clayton Stephenson, and YaleUniversity. The radiometric work waslargely supported by National ScienceFoundation grants BNS-9106787, BNS-8919236 to ALD. We further thank SallyMcBrearty and John Kingston for assistancein the field.

References

Best, M. G., Christiansen, E. H., Deino, A. L.,Gromme, C. S. & Tingey, D. G. (1995). Correlationand emplacement of a large, zoned, discontinuouslyexposed ash-flow sheet: 40Ar/39Ar chronology,paleomagnetism, and petrology of the PahranagatFormation, Nevada. J. geophys. Res. 100, 24593–24609.

Bishop, W. W. (1972). Stratigraphic succession ‘‘ver-sus’’ calibration in East Africa. In (W. W. Bishop &J. A. Miller, Eds) Calibration of Hominoid Evolution,pp. 219–246. Edinburgh: Scottish Academic Press.

Bromage, T. G., Schrenk, F. & Zonneveld, F. W.(1995). Paleoanthropology of the Malawi Rift: anearly hominid mandible from the Chiwoindo Beds,northern Malawi. J. hum. Evol. 28, 71–108.

Chapman, G. R. (1971). The geological evolution ofthe northern Kamasia Hills, Baringo District, Kenya.Ph.D. Dissertation, London University.

Chapman, G. R. & Brook, M. (1978). Chronostratig-raphy of the Baringo Basin, Kenya. In (W. W.Bishop, Ed.) Geological Background to Fossil Man.pp. 207–223. London: Geological Society ofLondon: Scottish Academic Press.

The KNM-BC 1 fossiliferous horizon liesapproximately in the middle of theChemeron Formation in the KapthurinRiver valley, encompassing the interval 3·3–1·6 Ma. This sequence is 0·5 Ma youngerthan the 5·3–3·8 Ma Chemeron Formationpreserved at Tabarin, 23 km to the north-northwest (Deino et al., 2002). Geologicreconnaissance has identified outcrops ofthe Chemeron Formation at Kipcherereand elsewhere that apparently bridge thisinterval.

Overall sedimentation rates measured forthe upper part of the formation along theKapthurin River (�11 cm/ka) are about halfthat determined for Tabarin (21–23 cm/ka).One possible explanation is that sedimenta-tion is occurring in a wedge-shaped trough,the deeper part of which is closest (e.g.,Tabarin) to the main north–south Saimofault. Another explanation may be thatthese results reflect waning tectonism by2·4–1·6 Ma.

Several other specimens potentiallyattributable to the genus Homo in the 2·4–2·0 Ma range have been reported in the lastdecade or so. Kimbel et al. (1996, 1997)provide reasonable evidence of Homo at2·33 Ma, from Hadar, Ethiopia. Schrenket al. (1993) and Bromage et al. (1995)describe a mandible from Malawi datedfaunally at 2·4 Ma, but their estimateappears to be based on a misunderstandingof data in Hill et al. (1992; see Hill, 1995,1999), and the specimen could be muchyounger. Suwa (1988, 1990) and Suwa et al.(1996) report isolated teeth from theShungura Formation, Omo, Ethiopia, thatcould plausibly belong to Homo datedbetween 2·4 Ma and 2·0 Ma. Yet the ageassignment of 2·456–2·393 Ma establishedherein for KNM-BC 1 makes it the oldestsecurely dated specimen attributed to genusHomo (Sherwood et al., 2001).

The lower date of 3·2 Ma provides ayounger age limit for the type specimensof Paracolobus chemeroni and Theropithecus

baringensis, both of which come from siteBPRP#97 (EAGRU JM90/91) at the base ofChemeron Formation in the KapthurinRiver.

151

Deino, A. & Potts, R. (1990). Single-crystal 40Ar/39Ardating of the Olorgesailie Formation, SouthernKenya Rift. J. geophys. Res. 95(B6), 8453–8470.

Deino, A. & Potts, R. (1992). Age probability spectrafor examination of single-crystal 40Ar/39Ar datingresults: example from Olorgesailie, southern KenyaRift. Quatern. Int. 7/8, 81–89.

Deino, A. L., Tauxe, L., Monaghan, M. & Hill, A.(2002). 40Ar/39Ar geochronology and paleomagneticstratigraphy of the Lukeino and lower ChemeronFormations at Tabarin and Kapcheberek, TugenHills, Kenya. J. hum. Evol. 1/2, 117–140.

Fuchs, V. (1934). The geological work of theCambridge Expedition to the East African Lakes,1930–31. Geol. Mag. 71, 97–112.

Fuchs, V. (1950). Pleistocene events in the BaringoBasin. Geol. Mag. 87, 149–174.

Gregory, J. W. (1896). The Great Rift Valley. London:John Murray.

Gregory, J. W. (1921). The Rift Valleys and Geology ofEast Africa. London: Seeley Service.

Hill, A. (1985). Early hominid from Baringo, Kenya.Nature 315, 222–224.

Hill, A. (1994). Late Miocene and early PlioceneHominoids from Africa. In (R. S. Corruccini & R. L.Ciochon, Eds) Integrative Paths to the Past: Paleo-anthropological Advances in Honor of F. Clark Howell,pp. 123–145. Engelwood Cliffs, NJ: Prentice Hall.

Hill, A. (1995). Faunal and environmental change inthe Neogene of east Africa: evidence from the TugenHills sequence, Baringo District, Kenya. In (E. S.Vrba, G. H. Denton, T. C. Partridge & L. H.Burckle, Eds) Paleoclimate and Evolution, withEmphasis on Human Origins. pp. 178–93. New Havenand London: Yale University Press.

Hill, A. (1999). The Baringo Basin, Kenya: from BillBishop to BPRP. In (P. Andrews & P. Banham, Eds)Late Cenozoic Environments and Hominid Evolution: aTribute to Bill Bishop, pp. 85–97. London: GeologicalSociety of London.

Hill, A. & Ward, S. (1988). Origin of the Hominidae: therecord of African large hominoid evolution between14 My and 4 My. Yearb. phys. Anthrop. 31, 49–83.

Hill, A., Drake, R., Tauxe, L., Monaghan, M., Barry,J. C., Behrensmeyer, A. K., Curtis, G., Fine Jacobs,B., Jacobs, L., Johnson, N. & Pilbeam, D. (1985).Neogene palaeontology and geochronology of theBaringo Basin, Kenya. J. hum. Evol. 14, 759–773.

Hill, A., Curtis, G. & Drake, R. (1986). Sedimentarystratigraphy of the Tugen Hills, Baringo, Kenya.In (L. E. Frostick, R. W. Renaut, I. Reid & J.-J.Tiercelin, Eds) Sedimentation in the African Rifts,pp. 285–295. Oxford: Blackwells & GeologicalSociety of London Special Publications 25.

Hill, A., Ward, S., Deino, A., Curtis, G. & Drake, R.(1992). Earliest Homo. Nature 335, 719–722.

Kimbel, W. H., Walter, R. C., Johanson, D. C., Reed,K. E., Aronson, J. L., Assefa, Z., Marean, C. W.,Eck, G. G., Bobe, R., Hovers, E., Rak, Y., Vondra,C., Yemane, T., York, D., Chen, Y., Evensen,N. M. & Smith, P. E. (1996). Late Pliocene Homoand Oldowan tools from the Hadar Formation

(Kadar Hadar Member), Ethiopia. J. hum. Evol. 31,549–561.

Kimbel, W. H., Johanson, D. C. & Rak, Y. (1997).Systematic assessment of a maxilla of Homo fromHadar, Ethiopia. Am. J. Phys. Anthropol. 103, 235–62.

Martyn, J. E. (1967). Pleistocene deposits and newfossil localities in Kenya. Nature 215, 476–480.

Martyn, J. E. (1969). The geological history of thecountry between Lake Baringo and the Kerio River,Baringo District, Kenya. Ph.D. Disseration, LondonUniversity.

McCall, G. J. H., Baker, B. H. & Walsh, J. (1967).Late Tertiary and Quaternary sediments of theKenya Rift Valley. In (W. W. Bishop & J. D. Clark,Eds) Background to Evolution in Africa, pp. 191–220.Chicago: Chicago University Press.

Pickford, M. (1975). Stratigraphy and palaeoecology offive late Cainozoic formations in the Kenya RiftValley. Ph.D. Dissertation, London University.

Pickford, M., Johanson, D. C., Lovejoy, O., White, T.& Aronson, J. L. (1983). A hominoid humeral frag-ment from the Pliocene of Kenya. Am. J. phys.Anthrop. 60, 337–346.

Renne, P. R., Swisher, C. C., Deino, A. L., Karner,D. B., Owens, T. L. & DePaolo, D. J. (1998). Inter-calibration of standards, absolute ages and uncertain-ties in 40Ar/39Ar dating. Chem. Geol. 145, 117–152.

Samson, S. D. & Alexander, E. C. Jr (1987). Calibra-tion of the interlaboratory 40Ar/39Ar dating standard,MMhb-1. Chem. Geol. Isot. Geosci. 6, 27–34.

Schrenk, F., Bromage, T. G., Betzler, C. G., Ring, U.& Juwayeyi, Y. M. (1993). Oldest Homo and Pliocenebiogeography of the Malawi Rift. Nature 365,833–836.

Sherwood, R. J., Ward, S. & Hill, A. (1996). Mandibu-lar fossa anatomy of the Chemeron temporal bone(KNM-BC 1). Am. J. Phys. Anthropol. Supplement22, 214–15.

Sherwood, R. J., Ward, S. C. & Hill, A. (2002). Thetaxonomic status of the Chemeron temporal(KNM-BC 1). J. hum. Evol. 1/2, 153–184.

Suwa, G. (1988). Evolution of the ‘‘robust’’ australo-pithecines in the Omo succession: evidence from man-dibular premolar morphology. In (F. E. Grine, Ed.)Evolutionary History of the ‘‘Robust’’ Australopithecines,pp. 199–222. New York: Aldine de Gruyter.

Suwa, G. (1990). A comparative analysis of hominiddental remains from the Shungura and Usno Forma-tions, Omo Valley Ethiopia. Ph.D. Dissertation,University of California, Berkeley.

Suwa, G., White, T. D. & Howell, F. C. (1996).Mandibular postcanine dentition from the ShunguraFormation, Ethiopia: crown morphology, taxonomicallocations, and Plio-Pleistocene hominid evolution.Am. J. phys. Anthrop. 101, 247–282.

Taylor, J. R. (1982). An Introduction to Error Analysis.Mill Valley, CA: University Science Books.

Tobias, P. V. (1967). Pleistocene deposits and newfossil localities in Kenya. Nature 215, 478–480.

Ward, S. & Hill, A. (1987). Pliocene hominid partialmandible from Tabarin, Baringo, Kenya. Am. J.phys. Anthrop. 72, 21–37.