The anomalous archaic Homo femur from Berg Aukas, Namibia

13 Pages·1999·0.3 MB·English

AMERICANJOURNALOFPHYSICALANTHROPOLOGY110:379–391(1999)
TheAnomalousArchaicHomoFemur From BergAukas,
Namibia:ABiomechanicalAssessment
ERIKTRINKAUS,1,2*CHRISTOPHERB.RUFF,3
ANDGLENNC.CONROY1,4
1DepartmentofAnthropology,WashingtonUniversity,St.Louis,
Missouri63130
2U.M.R.5809duC.N.R.S.,Laboratoired’Anthropologie,
Universite´deBordeauxI,33405Talence,France
3DepartmentofCellBiologyandAnatomy,JohnsHopkinsUniversity
SchoolofMedicine,Baltimore,Maryland21205
4DepartmentofAnatomyandNeurobiology,WashingtonUniversitySchool
ofMedicine,St.Louis,Missouri63110
KEYWORDS hominid;lowerlimb;biomechanics;Pleistocene
ABSTRACT The probably Middle Pleistocene human femur from Berg
Aukas,Namibia,whenorientedanatomicallyandanalyzedbiomechanically,
presentsanunusualcombinationofmorphologicalfeaturescomparedtoother
Pleistocene Homo femora. Its midshaft diaphyseal shape is similar to most
otherarchaicHomo,butitssubtrochantericshapealignsitmostcloselywith
earlierequatorialHomofemora.Ithasanunusuallylowneckshaftangle.Its
relative femoral head size is matched only by Neandertals with stocky
hyperarctic body proportions. Its diaphyseal robusticity is modest for a
Neandertal,butreasonablecomparedtoequatorialarchaicHomofemora.Its
gluteal tuberosity is relatively small. Given its derivation from a warm
climatic region, it is best interpreted as having had relatively linear body
proportions(affectingproximaldiaphysealproportions,shaftrobusticity,and
glutealtuberositysize)combinedwithanelevatedleveloflowerlimbloading
duringdevelopment(affectingfemoralheadsizeandneckshaftangle).AmJ
PhysAnthropol110:379–391,1999. r1999Wiley-Liss,Inc.
In1995,Grine,Jungers,TobiasandPear- earlyHomofemoraandalignsitwithmem-
sondescribedtheproximalhalfofaheavily bers of H. erectus and later Pleistocene
mineralized human femur, which had been Homo. The absence of a pilaster distin-
recovered30yearsearlierduringvanadium guishes it from early and recent modern
mining operations at the Berg Aukas mine humans,whereasthestrongdevelopmentof
in northern Namibia (Conroy et al., 1993; amedialbuttressplacesitclosesttoMiddle
Grineetal.,1995).Thespecimenisundated and Late Pleistocene archaic Homo femora
(Fig.1).
given its lack of stratigraphic association
Intheirassessmentofthefemur,Grineet
withanychronologicalindicators(biological
al. (1995) emphasized the large size of the
orgeological),butanassessmentofitsmor-
femoralhead,itsdiaphysealcross-sectional
phology relative to that of other hominid
femoraledGrineandcolleaguestoconclude
that it is best attributed to Middle or Late Grantsponsors:NationalScienceFoundation;C.N.R.S.;the
PleistocenearchaicHomo,aconclusionwith Wenner-Gren Foundation; the L.S.B. Leakey Foundation; the
NationalGeographicFoundation.
whichweconcur.Inparticular,themorphol- *Correspondenceto:ErikTrinkaus,DepartmentofAnthropol-
ogyoftheproximalepiphysealregiondistin- ogy,CampusBox1114,WashingtonUniversity,St.Louis,MO
63130,U.S.A.E-mail:[email protected]
guishes it from Australopithecus and some Received25November1998;accepted23June1999.
r1999WILEY-LISS,INC.
380 E.TRINKAUSETAL.
MATERIALS AND METHODS
Comparativesamples
Thebiomechanicallyrelevantparameters
of the BergAukas femur are compared pri-
marily to those of three chronological (and
bydefaultpartiallygeographical)samplesof
archaic Homo (see Tables 2 and 3 for speci-
mens).
The first sample, designated ‘‘Early Ar-
chaic Homo,’’ consists of mature femora
whicharedistinctfromthoseofAustralopi-
thecus,assignedtoearlyHomoandH.erec-
tus,anddatefromtheterminalPliocene(ca.
2.0maB.P.)totheearlyMiddlePleistocene
(ca. 600 ka B.P.). Additional data are in-
cludedfortheEarlyPleistoceneearlyadoles-
centKNM-WT15000specimen.Thissample
andperiodrepresentalongstageofstasisin
relativeendocranialcapacityandpelvicpro-
portions(Ruff,1995;Ruffetal.,1997);italso
includesonlyspecimensfromwarmclimate
zones. Since proximal femoral proportions
are influenced by pelvic configurations,
which in turn are affected by trends in
encephalization and ecogeographical pat-
ternsinbodyshape(Ruff,1995),thissample
shouldencompassspecimenswhosefemoral
variationreflectsprimarilylocomotorparam-
etersandnotseparateeffectsfromencepha-
lization,pelvicapertureshape,andrelative
Fig. 1. Anterior view of the Berg Aukas hominid
bodybreadth.
femur, taken relative to the anteversion plane of the
proximalfemur.Scaleincentimeters. The second sample, the one to which the
BergAukasfemurmayhavecloseaffinities,
is designated ‘‘Middle Archaic Homo.’’ It
properties, its low neck shaft angle, and consists of specimens from the middle por-
severaldetailsofthediaphysealshape.How- tion of the Middle Pleistocene, ca. 500 ka
ever, their assessment did not include a B.P.toca.200kaB.P.Femoralvariabilityin
biomechanical evaluation of the specimen. this group reflects both locomotor variation
On-going assessments of Pleistocene and and the marked encephalization that oc-
recent Homo femoral and tibial cross-sec- curred during the Middle Pleistocene (Ruff
tionalgeometry(e.g.,Ruffetal.,1993;Ruff, etal.,1997).Italsoincludesspecimensfrom
1995;Trinkaus,1997;Trinkausetal.,1999) acrosstheOldWorld(ZambiatoGermanyto
and proximal femoral proportions (Ruff et China),andsomeofitsvariationmayreflect
al., 1991; Trinkaus, 1993) have developed a human geographical and ecogeographic
frameworkwithinwhichwecannowassess variation.
the biomechanical implications of the Berg The third sample, ‘‘Late Archaic Homo,’’
Aukas femur. Given the dearth of Middle includes the Neandertals sensu lato from
Pleistocene human postcrania, any insight the terminal Middle Pleistocene to the
which this, albeit undated, specimen might middleofoxygenisotopestage3(ca.150ka
provide would add to our knowledge of ge- B.P.toca.35kaB.P.).Thissample,theonly
nusHomolocomotorevolution. oneoflatearchaichumanswhichpreserves
BERGAUKASHOMOFEMUR 381
femora, is notable for both its encephaliza- respecttotheirbodyproportions(Ruff,1991,
tion approaching that of recent humans 1994;Holliday,1997a,b).Itisthereforenec-
(Ruff et al., 1997) and its hyperarctic body essary to know the body breadth of the
proportions(Holliday,1997b). individuals in question or to estimate it
In addition, a Eurasian sample of early based on ecogeographical patterns, or to
modern humans ($ 18 ka B.P.) is used to take known variation in relative body
providearobustmodernhumanbaselinefor breadthintoconsiderationinassessmentsof
the assessments of archaic Homo femora. proximalanddiaphysealdimensionstofemo-
These specimens derive from the sites of rallength.Giventheincompletestateofthe
AreneCandide,BarmaGrande,Caviglione, Berg Aukas femur, the latter approach is
Cro-Magnon, Doln´ı Veˇstonice I and II, takenhere.
Grotte-des-Enfants, Minatogawa, Mladecˇ,
The assessment of subtrochanteric (80%)
NahalEinGev,OhaloII,Paglicci,Parabita,
diaphysealshapewasdoneusingmaximum
Paviland, Pavlov I, Prˇedmost´ı, Qafzeh, La
to minimum second moments of area (I
Rochette, Skhul, and Willendorf (N 5 16 to max
andI ).Givenvariationintheorientation
34forvariouscomparisons). min
ofthemajoraxisatmidshaft,frommediolat-
eral in EarlyArchaic Homo to anteroposte-
Methods
rior in early modern humans, the anatomi-
Six biomechanical parameters of interest
callyorientedI andI valuesareemployed
x y
are preserved on the Berg Aukas femur:
atthatlevel.ForI andI ,thesagittalplane
x y
head size reflecting hip joint reaction force
of the femur is defined as passing through
levelsandhencebodymass(Jungers,1988;
the linea aspera and the mediolateral mid-
McHenry, 1988; Ruff et al., 1991), femoral
pointofthediaphysisnearmidshaft.
neck length influencing gluteal abductor
Thecross-sectionalgeometricparameters
musclemomentarmsandloaddistributions
(see Table 1 for a list) for the comparative
through the proximal femur (Lovejoy et al.,
samples were computed using versions of
1973; Ruff, 1995), gluteal tuberosity size
SLICE(NagurkaandHayes,1980).Mostof
relatedtohipmusclehypertrophy(Trinkaus,
the cross sections were obtained noninva-
1976),diaphysealstrengthreflectingoverall
sivelybymoldingthesubperiostealcontour,
locomotor levels (Ruff et al., 1993), subtro-
determiningcorticalthicknessesusingmulti-
chanteric diaphyseal shape reflecting hip
or biplanar radiography, and interpolating
region proportions and the resultant rela-
theendostealcontourwithintheboundaries
tiveanteroposteriorversusmediolateralload
set by the cortical thicknesses. A few were
levels(Ruff,1995),andmidshaftdiaphyseal
fromscaledphotographsofappropriatefos-
shape reflecting in part hip proportions but
silization breaks, and several were taken
mainly locomotor levels (Ruff, 1987, 1995,
1999; Trinkaus et al., 1998). The first four frompublishedcross-sections(e.g.,Weiden-
require the scaling of the relevant femoral reich,1941;Mallegnietal.,1983).
measure to body mass or body mass times Anteroposterior head diameter was mea-
beam length. The second two involve com- sured directly or, for OH-28, Arago 44 and
parisons of relevant perpendicular second Amud 1, was estimated from acetabular
momentsofarea. height using a regression based on recent
The scaling of head size, neck length, humans(r250.896).Necklengthwasquan-
glutealtuberositysize,andmidshaftdiaphy- tifiedusingthebiomechanicalnecklengthof
sealrigidityrequiretheestimationoffemo- Lovejoy et al. (1973), the distance perpen-
ral length, since it approximates the beam diculartothediaphysealaxisfromthemost
lengthforthefemurandisthebestestima- lateralpointofthegreatertrochantertoits
torofstature(Feldesmanetal.,1990).Body tangenttotheproximalfemoralheadtaken
massisdependentuponstatureandrelative inthecoronalplaneofthefemoralheadand
body breadth (Ruff, 1994; Ruff et al., 1997), neck.Glutealtuberositysizewasquantified
and both recent and Pleistocene Homo are asthemaximumbreadthoftherugosearea
knowntohavevariedecogeographicallywith ofthetuberosityproper(Trinkaus,1976).
382 E.TRINKAUSETAL.
TABLE1. Cross-sectionalgeometricpropertiesofthemidshaft(50%level)
andsubtrochantericlevel(80%)oftheBergAukasfemur1
50% 50% 50%cast 80%Grineetal. 80%
Grineetal.(1995) CTscan (adjusted) (1995) adjusted
TA 853 800 805 898 842
CA 750 708 697 794 744
I — 51,599 50,365 — —
x
I — 51,050 52,720 — —
y
I 67,149 61,019 59,889 84,677 75,604
max
I 49,112 41,631 43,917 47,362 42,288
min
J 116,261 102,649 103,085 132,038 117,892
Theta — 134° 139° — —
1Areasinmm2andsecondmomentsofareainmm4.TA,totalarea;CA,corticalarea;Ix,anteroposteriorsecondmomentofarea;Iy,
mediolateralsecondmomentofarea;Imax,maximumsecondmomentofarea;Imin,secondmomentareaperpendiculartoImax;J,polar
momentofarea;theta:orientationofImax.
Assessments of the biomechanical affini- gluteal tuberosity on the medial diaphysis
ties of the Berg Aukas femur relative to and gradually rotates posteriorly as it goes
thoseofotherarchaicHomospecimenswere distally (Trinkaus, 1976, 1984). On femora
donegraphicallyforallbutthenecklength where it is well developed, the medial but-
tobonelengthcomparison(Figs.3–6).Since tress reaches the posteromedial aspect of
all of these variables are, or may well be, the diaphysis near midshaft. On the Berg
relatedtoeachotherinnonlinearways(Ruff Aukas femur, this buttress is directly pos-
et al., 1993), the data were logged prior to teromedialatthedistalbreakoftheshaft.It
graphingandanalysis.Inaddition,az-score wasthereforeassumedthatthedistalbreak
from the early modern human reduced ma- isclosetomidshaft(seeGrineetal.,1995).
jor axis regression line is provided for each The distance parallel to the diaphyseal
archaicHomofemur. axis, from the proximal neck to the distal
breakisca.235mm,providingabiomechan-
BergAukasfemur
icallength(RuffandHayes,1983)ofca.470
Thisanalysisisbasedondataderivedfor mm. Given that the medial buttress had
the Berg Aukas femur, combining the data alreadyreachedtheposteromedialaspectof
presentedbyGrineetal.(1995)(headdiam- theshaft10to15mmproximalofthedistal
eter, neck length, and neck shaft angle), break,thisestimateismorelikelytooveres-
data derived from computed tomographic timatethanunderestimatethebone’sorigi-
(CT) scans of the specimen by GCC (taken nal length. The addition of the 10 mm (the
using a Siemens Somatom DR3 scanner at distancefromtheproximalnecktotheproxi-
125 kV with a window width of 2050 and mal head measured parallel to the diaphy-
slice thickness of 1mm) (for cross-sections) sealaxis)providesamaximumlengthca.480
andahighqualityresincastofthespecimen mm,38mmbelowthepreviousmeanpredic-
(for50%cross-sectionsandglutealtuberos- tionof518mm.
itybreadth). To obtain I and I as well as the other
x y
Grineetal.(1995,pp.159–160)estimated cross-sectional parameters, the bone was
lengthusingamultipleregressionbasedon oriented using the linea aspera, the distal
a recent human (AfricanAmerican) sample breakofthecastwasphotographed(Fig.2),
andsixmeasurementsoftheheadandneck. the image was projected onto a Summa-
This provided an ‘‘interarticular’’ length of graphics digitizing tablet enlarged linearly
518 6 16 mm. To avoid any circularity in 5.95times,andthesubperiostealandendos-
assessing proximal epiphyseal proportions teal contours were digitized (not including
in the present analysis, the midshaft of the the trabeculae in the posterior medullary
Berg Aukas femur was determined by the cavity) (Fig. 2). The same procedure was
relative position of the medial buttress, a followedforthehardcopyprint-outofourCT
swelling of bone along the medial proximal image taken just proximal of the distal
diaphysis which begins at the level of mid- break, linearly enlarged 3.84 times. Cross-
BERGAUKASHOMOFEMUR 383
RESULTS
Overallsize
The Berg Aukas femur is very large. Its
femoral head diameter of 57.6mm is the
largestknown,althoughfemoralheaddiam-
etersbetween50and55mmareknownfor
several Middle and Late Pleistocene hu-
mans,includingArago44,BrokenHillE689
and E907, Krapina 213, La Chapelle-aux-
Saints1,LaFerrassie1,Neandertal1,and
Spy 2 (only the largest male early modern
humans have femoral head diameters ex-
ceeding50mm).Itsmidshaftcorticalareaof
697mm2isthelargestknownforanarchaic
Homo femur, even though it is approached
by those of KNM-ER 736 (659 mm2),
Ehringsdorf 5 (666 mm2), andAmud 1 (659
mm2).Itsmidshaftpolarmomentofareaof
103,085mm4issimilarlyamongthelargest
known, being surpassed only by KNM-ER
Fig.2. DistalviewofacastoftheBergAukasright 736(116,628mm4)andmatchedbythoseof
hominidfemur,showingthecross-sectionaldistribution
ofbonerevealedbythetransversebreakofthediaphy- Ehringsdorf5(102,449mm4),andQafzeh8
sis.Anterior is above and medial is to the right. The (102,428mm4).
pencilmarkindicatesthepositionofthelineaaspera.
These large femoral dimensions and the
Scaleinmillimeters.
body mass estimate based on femoral head
diameterofca.93kg(Grineetal.,1995)are
impressive.Theyalsoemphasizetheneedto
sectionalparameterswerecomputedusinga
scaletheseparameterstoappropriatemea-
PC-DOSversion(Eschman,1992)ofSLICE
suresofbodysize.
(Nagurka and Hayes, 1980). Each image
was digitized twice and the results aver-
Hipproportions
aged. The resultant values were then ad-
justed for linear shrinkage (ca.1.6%) of the There are few archaic Homo specimens
cast(Table.1).Thefinalvaluesarelessthan thatprovidebiomechanicalnecklengthand
those calculated by Grine et al. (1995), but femoral length (or reasonable estimates of
they remain consistent with the external them),anditisinappropriatetoscalefemo-
diametersofthediaphysis.1 ralnecklengthagainstotherproximalfemo-
Forcomparable80%measures,thevalues ral dimensions since they are frequently
provided by Grine et al. (1995) were ad- highlycorrelated(Wolpoff,1978;Ruff,1995).
justed by the percentage differences found All of the other archaic Homo femora fall
for the midshaft cross sections (6.7% and abovetheregressionlinefortheearlymod-
12.0%) (Table 1). Since these are linear ernhumansample(Table2),withKNM-ER
correctionsandtheonlysubtrochantericval- 1481a and Spy 2 having relatively high
uesofinteresthereareImaxandImin,thishas values. In addition, the early H. erectus
atrivialeffectonthecomparisons. KNM-WT15000earlyadolescentprovidesa
z-score of 2.93, in agreement with previous
assessmentofitsfemoralnecklength(Ruff,
1Toassesswhichvaluesaremorereasonable,totalsubperios- 1995).ArchaicHomotendtohaverelatively
tealareaswereregressedontheproductofthemidshaftdiam-
eters (‘‘area’’) for archaic Homo femora (50% TA 5 0.705 3 longfemoralnecks(seealsoWolpoff,1978).
‘‘area’’18.33,r250.953,N531;BergAukasAP:34.0mm,ML:
TheBergAukasfemurisinthemiddleof
33.5mm),andz-scoresforthedifferentBergAukastotalarea
residualswerecomputed.TheBergAukasz-scoresare1.34for theearlymodernhumansampleandhasthe
thetotalareafromGrineetal.(1995),20.45fortheCTscan
digitizedhere,and20.27fortheshrinkage-adjustedcastvalues. relativelyshortestneckofthearchaicHomo
384 E.TRINKAUSETAL.
TABLE2. z-Scoresforcomparisonsoffemoralnecklength,headdiameter,andglutealtuberositybreadth
Necklength/ HeadDiameter/ Glutealtuberosity/ Glutealtuberosity/
length length length head
BergAukas 0.06 2.48 21.02 24.88
EarlyArchaicHomo
KNM-ER737 — — 2.51 —
KNM-ER1472 1.27 20.36 — —
KNM-ER1481a 2.00 1.65 4.44 2.17
GesherBenotYa’aqov1 — — 4.27 —
OH28 — 0.09 1.63 1.21
MiddleArchaicHomo
BrokenHillE689 — — — 21.01
BrokenHillE690 — — 0.82 —
BrokenHillE907 — — — 21.85
LateArchaicHomo
Amud1 — 20.26 0.79 0.89
LaChapelle1 0.60 3.09 2.30 22.02
LaFerrassie1 1.09 2.34 2.79 20.07
LaFerrassie2 0.40 1.70 4.92 2.86
Krapina213 — — — 0.30
Krapina214 — — — 4.01
Neandertal1 0.95 2.65 2.90 20.60
St.-Ce´saire1 — — 3.02 —
Shanidar1 — — 1.62 —
Shanidar4 — 2.11 4.09 1.46
Shanidar5 — 0.52 — —
Shanidar6 — — 5.48 —
Spy2 2.58 3.59 3.09 21.68
Tabun1 0.40 0.92 3.91 2.67
femora.Aswiththecomparisonofanatomi- have platymeric proximal femoral diaphy-
calnecklengthtoitsverticaldiameter(Grine ses, reflecting the biomechanical loads im-
et al., 1995), the Berg Aukas femoral neck posed on their femora from long femoral
length is unexceptional from a modern hu- necksandplatypelloidpelves.TheEurasian
man perspective, although relatively short Late Archaic Homo sample has relatively
foranarchaicHomo. roundproximalfemoraldiaphyses,withthe
Thismoderatenecklength(foranarchaic two exceptions being the early (terminal
Homo) is associated with the lowest known
Middle Pleistocene) Krapina specimens.
(presumably)nonpathologicalhumanfemo-
Most of the early modern humans exhibit
ralneckshaftangle.Itsvalueof106°(Grine
platymericfemora,withtheexceptionofthe
etal.,1995,p.163)isbelowallotherknown
very linear Qafzeh-Skhul sample. This is
values for archaic Homo (being approached
reflectedintheirz-scores,whicharepositive
by KNM-WT 15000 at 110° andAmud 1 at
to minimally negative for the early sample
113°) (Walker and Leakey, 1993; Trinkaus,
andnegative(exceptfortheKrapinafemora)
1993) and 3.4 and 4.1 standard deviations
forthelatearchaicsample(Table3).
below the means of two Holocene human
The Middle Pleistocene remains exhibit
groups with the lowest mean neck shaft
considerable variability. Two Broken Hill
angles [Khoisan males (Grine et al., 1995)
femora are quite round, one is relatively
and Japanese Jomon foragers (Ishisawa,
platymeric,andtheTabunE1femurisclose
1931), respectively] and below a minimum
to the early modern human line. The Berg
value of 110° for a sample of 1376 recent
humansfrom15populations(Andersonand Aukas femur is in the middle of the early
Trinkaus,1998). modern human and overall distributions
(Table 3). It is therefore relatively platy-
Femoraldiaphysealshape
meric, more similar to the Early Archaic
The distribution of 80% I versus I Homo sample than it is to the LateArchaic
man min
(Fig.3)reflectsthepatternpreviouslynoted Homogroup,fallingveryclosetospecimens
(Ruff, 1995; Trinkaus et al., 1998; Trinkaus suchasKNM-ER736and803aandGesher
andRuff,1999).Inthis,EarlyArchaicHomo BenotYa’aqov1.However,itfitswellwithin
BERGAUKASHOMOFEMUR 385
eral to posteromedial, as is reflected in its
thetavalueof139°,theanteroposteriorand
mediolateral second moments of area (I
x
versusI )areverysimilar(Table1).
y
With the exceptions of the Broken Hill
E690andCasteldiGuido1femora,allofthe
archaic Homo femora fall on the rounded
sideoftheearlymodernhumandistribution
(Fig. 3). The mean z-scores for the Early,
MiddleandLateArchaicHomosamplesare
23.31, 22.44 (22.96 without Broken Hill
E690 and Castel di Guido 1), and 22.85,
respectively. The Berg Aukas femur with a
z-scoreof23.51fallswellwithintheranges
of archaic Homo femora and outside of the
range of the largely pilastric early modern
humanfemora.
Femoraldiaphysealrobusticity
Even though the Berg Aukas femur is
large,itslevelofrobusticity(strengthscaled
to appropriate measures of body size and
limb length) is less apparent relative to
other Pleistocene Homo femora. This ambi-
guity derives from the combined contribu-
tions of body proportions, body mass and
activity level to the hypertrophy of the
weight-bearing locomotor skeleton (Ruff et
al.,1993).
Fig.3. Bivariatedistributionsofperpendicularsec- In comparisons of both midshaft cortical
ondmomentsofareaforthesubtrochanteric(80%)level areasandpolarmomentsofareatofemoral
(above)andthemidshaft(50%)level(below).Thesubtro-
chanteric distribution employs maximum and perpen- lengths, there is a consistent pattern. Late
dicular to maximum second moments of area (I Archaic Homo specimens are generally on
max
versusI ),whereasthemidshaftoneplotsanteroposte-
min the high side of the early modern human
rior versus mediolateral ones (I versus I). Symbols:
solidhexagon,BergAukasfemur;xgraydiamyonds,early distribution,withmeanz-scoresof1.49and
archaic Homo; dark gray pentagons, middle archaic 1.53(Fig.4andTable3).TheEarlyArchaic
Homo;opensquares,latearchaicHomo;opentriangles,
Homo femora plus the Middle Pleistocene
earlymodernhumans.Thereducedmajoraxislinefor
theearlymodernhumansampleisprovided. BrokenHillE690femur,allofwhichderive
from Africa or the neighboring Near East,
have more modest levels of relative cortical
thevariabledistributionoftheMiddlePleis- area and polar moments of area, being
tocenefemora. slightly higher on average than the early
Inthemidshaft,BergAukasfallswiththe modern human sample (mean z-scores of
majority of the archaic Homo specimens. It 0.52and0.36,respectively).
lacksanyevidenceofapilaster,sinceitonly BergAukas has a z-score of 1.39 for rela-
has a slight projection of the linea aspera tive cortical area, well within the Late Ar-
dorsally (Fig. 2). There is a marked medial chaicHomoplusZhoukoudianrange(mean:
buttress, one of the largest known for an 1.57)butoverlappingthoseofA¨ınMaarouf1
archaic Homo femur, being approached by and KNM-ER 803a. It has a z-score of 1.00
those ofAmud 1, La Ferrassie 1 and Saint- for the relative polar moment of area, be-
Ce´saire 1 (Trinkaus et al., 1991, 1998; tween the mean values for the Early (plus
Trinkaus and Ruff, 1999). Therefore, even Broken Hill; 0.36) and Late (plus Zhoukou-
thoughthemajoraxisisorientedanterolat- dian; 1.59) Archaic Homo samples. Is Berg
386 E.TRINKAUSETAL.
TABLE3. z-Scoresforcomparisonsoffemoralcross-sectionalproperties
80%I /I 50%I/I 50%CA/Len 50%J/Len
max min x y
BergAukas 20.10 23.51 1.39 1.00
EarlyArchaicHomo
A¨ınMaarouf1 — 21.75 1.73 1.42
KNM-ER736 20.26 23.81 (20.78) (21.11)
KNM-ER737 1.06 24.54 0.34 0.57
KNM-ER803a 20.02 23.71 1.60 1.72
KNM-ER1472 0.96 21.91 1.06 0.51
KNM-ER1481a 0.42 22.74 0.19 20.44
KNM-ER1808mn 2.17 23.46 (0.53) (1.44)
GesherBenotYa’aqov1 20.03 22.87 0.14 20.47
OH28 1.30 25.02 20.39 20.01
MiddleArchaicHomo
BrokenHillE689 22.04 — — —
BrokenHillE690 1.23 20.23 (0.76) (20.08)
BrokenHillE793 — 21.81 — —
BrokenHillE907 21.73 — — —
CasteldiGuido1 — 20.53 — —
Ehringsdorf5 — 22.91 — —
TabunE1 20.26 23.02 — —
ZhoukoudianF1 — 23.57 (2.50) (2.25)
ZhoukoudianF2 — 23.35 — —
ZhoukoudianF4 — 23.39 — —
ZhoukoudianF5 — 23.41 — —
ZhoukoudianF6 — 22.20 — —
LateArchaicHomo
Amud1 20.55 23.29 1.40 0.93
LaChapelle1 22.57 23.45 2.20 2.08
LaFerrassie1 24.11 23.89 0.60 0.56
LaFerrassie2 21.41 23.35 1.35 1.80
Fond-de-Foreˆt1 — 22.91 1.01 1.13
Krapina213 0.31 — — —
Krapina214 0.48 — — —
Neandertal1 22.04 21.99 0.95 1.08
St.-Ce´saire1 21.74 21.67 2.33 2.29
Shanidar4 — 22.87 2.51 2.87
Shanidar5 — 23.10 1.89 2.32
Shanidar6 — 22.39 1.61 1.95
Spy2 21.95 22.89 1.09 1.01
Tabun1 21.00 23.58 0.93 0.39
Tabun3 — 21.62 — —
Aukas one of the more robust earlier Pleis- the comparative samples, an assumption
tocene archaic Homo femora, similar to a knowntobefalse.
modestly built Late Pleistocene archaic Such a comparison for Pleistocene Homo
Homo, or can its relative diaphyseal hyper- placesBergAukasatthetopofthedistribu-
trophybeotherwiseexplained? tion, exceeded in relative femoral head size
onlybythreeEuropeanNeandertals(Fig.5
Relativefemoralheadsize
andTable2).Itsz-scoreof2.48isclosetothe
To assess the relative size of the Berg mean (2.67) of the European Late Archaic
Aukasfemoralhead,itneedstobescaledto Homosampleandwellabovethemoremod-
a measure of body mass. However, the only estvaluesformostofthelateNearEastern
possibleindicationsofbodymassareeither specimensandthethreeEarlyArchaicHomo
thefemoralheaditselforcalculationsbased specimens. The Berg Aukas femur has a
onestimatedstature(fromitsreconstructed large femoral head, but not one which is
length) and body breadth (from ecogeo- exceptional relative to at least one Pleis-
graphicalpatterninggivenitsinferredpaleo- tocenehominidsample.
climatological context) (Ruff et al., 1997).
Relativeglutealtuberositysize
Alternatively,itcanbecomparedtofemoral
length,bearinginmindthatsuchacompari- The gluteal tuberosities of many Pleis-
sonassumessimilarbodyproportionsacross tocene archaic and early modern humans
BERGAUKASHOMOFEMUR 387
Fig.5. Bivariatedistributionoffemoralheadantero-
posteriordiameterversusfemorallength.Symbolsasin
Fig.3.Thereducedmajoraxislinefortheearlymodern
humansampleisprovided.
ofM.gluteusmaximushypertrophy(andby
extensionthatofthehipmusculaturegener-
ally, given the need for balance between
synergistsandantagonists).
Gluteal tuberosity breadth relative to
femoral length largely separates archaic
Homoandearlymodernhumanfemora;the
onlyoverlapinvolvesthemodesttuberosity
dimensions for Broken Hill E690 (possibly
underestimated given surface abrasion to
thespecimen)andLaChapelle-aux-Saints1
Fig. 4. Bivariate distributions of midshaft (50%) and the pronounced one for Doln´ıVeˇstonice
corticalarea(above)andpolarmomentofarea(below)
13(Fig.6).ItplacesBergAukasamongthe
versusfemoralbiomechanicallength.SymbolsasinFig.
3. The reduced major axis line for the early modern moregracileofthespecimens,withaz-score
humansampleisprovided. of20.02.Analternativeapproachistocom-
pare gluteal tuberosity breadth to femoral
head diameter, assuming that the latter is
are large and rugose (Trinkaus, 1976). The better reflecting body mass than is femoral
BergAukas tuberosity is similarly marked, length across these ecogeographically vari-
descending from a third trochanter just be- ablesamples(Fig.6).Theresultisageneral
low the level of the lesser trochanter and uniformity of early modern human and ar-
becoming increasingly rugose and concave chaic Homo femora, with only three small
asitdescendsdistallyontothedorsalproxi- Late Archaic females (La Ferrassie 2,
mal diaphysis. Its maximum breadth of 9.0 Krapina 214 and Tabun 1) plus the early
mmislocatedproximally,atthedistalendof KNM-ER1481abeingrobustoutliers(Fig.6
thethirdtrochanter,andthetuberositynar- and Table 2). Berg Aukas, with its modest
rowstoca.7.5mmdistally. glutealtuberositybreadthandlargefemoral
Eventhoughitisdifficulttoassessactual head,isanunusuallylowoutlier.
muscle size from osteological attachment
DISCUSSION
dimensionsandonlyaportionofM.gluteus
Paleoenvironmentalcontext
maximusinsertsintotheglutealtuberosity
(the remainder blending into the iliotibial The Pleistocene paleoenvironments of
tract), the size of the gluteal tuberosity southwestern Africa are poorly known, but
appears to provide a reasonable indication itappearsfromgeologicalevidence(Wardet
388 E.TRINKAUSETAL.
current arid to extreme-arid environment
progressivelydevelopedinconjunctionwith
the late Miocene origin of the Benguela
current.
The scattered Middle Pleistocene archeo-
logical evidence and geological evidence for
at least ephemeral lakes (e.g., Shackley,
1980; Sandelowsky, 1983), as well as the
presence of the BergAukas hominid femur,
indicate that there were periods during
which rainfall was sufficient to support a
savannagrasslandandassociatedfauna.
WecanonlyassumethattheBergAukas
hominid was associated with the post-
Miocene(probablyPleistocene)micro-mam-
malian fauna represented in one of the
clusters of breccia blocks from the Berg
Aukas mine dump (Conroy, 1996). Ten ex-
tant mammalian genera are exclusive to
thiscluster,themajorityofwhichareassoci-
ated with semi-arid to desert conditions
(SkinnerandSmithers,1990).
Itisthereforelikelythatthepaleoenviron-
mental context of the Berg Aukas hominid
was similar to that of modern Namibia,
warmandsemi-aridtoarid.Alinearhuman
body form, such as is associated with such
environments (Ruff, 1994), appears most
likelyforthisindividualbasedonitspaleo-
Fig.6. Bivariatedistributionsofglutealtuberosity environmentalcontext.
breadth versus femoral biomechanical length (above)
andfemoralheaddiameter(below).SymbolsasinFig.3.
Thereducedmajoraxislinefortheearlymodernhuman Diaphysealshape
sampleisprovided.
Thesubtrochantericandmidshaftdiaphy-
sealcrosssectionsoftheBergAukasfemur
fit best within the variable Middle Archaic
al., 1983) that it remained relatively warm
Homo sample. The specimen is fully non-
in the Pleistocene and that Namibia was
pilastric, and its midshaft structural distri-
semi-arid to arid throughout this period. A
bution of bone places it close to most of the
serious limitation to Pleistocene paleocli-
other archaic Homo femora (all except two
maticreconstructionsofNamibiaisthepau-
unusual Middle Pleistocene specimens).
city of time successive mid-to-late tertiary
ProximallyitislessroundedthantheNean-
vertebrate fossil localities (Shackley, 1980;
Senut et al., 1992; Pickford et al., 1994). dertals and closer to the earlier (and lower
This problem has been alleviated partially latitude) specimens. However, it is unclear
by the discovery of a dozen fossiliferous whetherthetrendtowardrounderproximal
localitiesinnorthernNamibia,ofwhichthe femoral diaphyses through later archaic
richest,bothintermsoftaxonomicdiversity Homo was the result of changing load pat-
andspecimenabundance,istheBergAukas terns in this region from trends in pelvic
mine (Senut et al., 1992; Conroy, 1996). In apertureshape(Ruff,1995)orwasacombi-
general, the biostratigraphic evidence at nation of such aperture changes combined
BergAukaslargelysupportsotherevidence with ecogeographically influenced pattern-
(Siesser, 1980) suggesting that Namibia’s inginpelvicbreadth(Ruff,1991).