Biological Adaptations and Affinities of Bronze Age Harappans

48 Pages·2008·22.61 MB·English

Harappa Excavations
1986-1990
A Multidisciplinary Approach to
Third Millennium Urbanism
Edited by Richard H. Meadow
Monographs in World Archaeology No.3
• PREHISTORY PRESS
Madison Wisconsin
Copyright©1991bytheIndividualAuthors.
PrehistoryPress
7530WestwardWay
Madison,Wisconsin53717-2009
JamesA. Knight,Publisher
CarolJ. Bracewell,ManagingEditor
AllRightsReserved
Manufactured in theU.S.A.
ISBN0-9629110-1-1
ISSN1055-2316
LibraryofCongressCataloging-in-PublicationData
Harappaexcavations1986-1990 : amultidisciplinaryapproach to third millenniumurbanism /
editedbyRichard H.Meadow.
p. em. - - (Monographsin worldarchaeology,ISSN1055-2316 : no.3)
Includesbibliographicalreferences.
ISBN0-9629110-1-1 : $33.00
1.HarappaSite(Pakistan) 2.Excavations(Archaeology)
- Pakistan. I.Meadow,RichardH. II. Series.
05392.2H3H37 1991
934--dc20 91-39504
CIP
Coverart: Bowl on Stand H88-1002/192-17 associated with Burial 194a
in HarappanPhase Cemetery (see Figure 13.18).
11
Biological Adaptations and Affinities
of Bronze Age Harappans
Brian E. Hemphilr, John R. Lukacs2 K.A.R. Kennedyl
,
IDivision ofBiological Sciences, Cornell University
2Department ofAnthropology, University ofOregon
Cranial and dental metric and non-metric data derived from human skeletal remains ofthe Harappan phase at
Harappa areemployedtogetherwith comparabledatafrom South Asiaand neighboringareas toaddress questions of
biologicaladaptationandcontinuity.Aprogressivedeclineindentalhealthisdemonstratedtohaveoccurredwithinthe
IndusValleyfrom theneolithicto the urbanphaseoftheHarappan Civilization. This accords wellwithexpectationsof
theproximateeffectsofincreasingrelianceuponagriculture.Sexualdifferencesindentalhealthanddifferences intooth
sizebetween humansfrom South Asiansitesofdifferent periods canalso berelated todiet andfood preparationtech
niques.AsforthequestionofbiologicalcontinuitywithintheIndusValley,twodiscontinuitiesappeartoexist.Thefirst
occurs between6000 and 4500 BCandis reflected bythe strongseparation in dental non-metric characters between
neolithicandchalcolithicburialsatMehrgarh. Thesecondoccursatsomepointafter800BCbutbefore200Be.In the
intervening period, while there is dental non-metric, craniometric, and cranial non-metric evidence for adegree of
internalbiologicalcontinuity,statisticalevaluationofcranialdata revealsclearindicationsofinteractionwiththeWest
andspecificallywiththeIranian Plateau.
T his paper seeks to employ biological data 1975;Meadow 1989).Theearliestevidencefor settled
derived from human skeletal remains recov agriculture in South Asia comes from aceramic and
ered from the Harappan phase cemetery laterneolithiclevelsatMehrgarh(ca. 6000BC),where
("CemeteryR37") atHarappa toaddress two current impressionsin mud-brick ofsix-row barley, einkorn,
areasofdiscussion raised byanalysesofnon-skeletal emmer, and durum-bread wheat were identified by
remains. These areas are progressive agricultural Costantini (1984). In the chalcolithic period (ca. 4500
intensificationandbiologicalcontinuity. BC), barley continued to be used but was accompa
The first oftheseareasconcernsbiologicaladapta nied by a dramatic increase in wheat utilization.
tion.Perhapsthemostsignificantadaptationsfacedby Costantini (1979, 1984, 1990), Jarrige (1985), and
Bronze Age Harappans were those incurred through Meadow (1989) have interpreted archaeobotanical
dietary changes wrought by increasing dependence remainsfrom the occupational sequencein the Kachi
on domesticated foodstuffs. Earlier workers consis Plain (Mehrgarh-Nausharo-Pirak) as comprising an
tently emphasized the agricultural economy of early(neolithic)barley/wheatsubsistencebaseestab
Harappansociety.Infactsomescholars,impressedby lishedbythemid-sixthmillenniumBCthatcametobe
the worker's platforms and granaries found in the supplemented in theearly 2nd millenniumby use of
northernpartofthesiteofHarappa(MoundF),devel othercultigenssuchasrice,millets,andsorghum.
oped elaborate interpretations of cultural Apartfrom thisdirectevidenceofcultivatedfood
developmentthathingedonforcesofeconomicdevel stuffs, many important technological developments
opment(Childe1950,1957). relating to food storage and preparation are also
Archaeological evidence ofincreasing agricultural evident from aceramic and chalcolithic levels at
intensification within the Indus Valley has been well Mehrgarh. These include the development of utili
documented (Allchin and Allchin 1982; Fairservis tarian pottery, grindstones, composite microlithic
137
138 Harappa Excavations 1986-1990: A Multidisciplinary Approach
sickles, and grain storage structures (Jarrige 1981, prevalence patterns (Walker and Hewlett 1990). In
1985; Jarrige and Meadow 1980; Jarrige and addition to these cultural factors, recent studies by
Lechevallier 1979; Lechevallier and Quivron 1981, Hildeboldt et a1. (1988, 1989) demonstrate that the
1985; Lechevallier et a1. 1982). Similarly, faunal distinctive bio-geochemical features of localized
remainsdemonstrateashiftfromwildtodomesticated geographicareas also influencedental disease preva
varietiesofcattle,goat,andsheepduringthecourseof lence. Together, these behavioral and geographic
the neolithic period (Jarrige and Meadow 1980; factors are responsiblefor thewell-documentedinter
Meadow1982,1984,1987,1989). populationaldifferencesinoralhealththataccompany
Evidence of increased dependence on agricultural changesindietarybehavior.
foodstuffsisabundantfromtheIndusCivilizationsites Although odontometric and dental pathology data
of Harappa, Kalibangan, and Mohenjo-daro (Allchin are often employed in other areas of the world
and Allchin 1982). Archaeological evidence indicates (Armelagos 1969; y'Edynak and Fleisch 1983; Frayer
thattwo varietiesofwheat, barley, field peas, sesame, 1987, 1989; Greene 1972; Hodges 1987; Larsen 1984;
and mustard were cultivated at these sites, while Machiarelli 1989; Molnar and Molnar 1985; Powell
ploughed fields during pre-Harappan times at 1988; Turner 1979), results of these studies are not
Kalibangan (Thapar 1973, 1975) suggest considerable consistent, and no comprehensive analysis of dental
antiquity for the contemporary practice of ploughing pathologyamongprehistoricpopulationsoftheIndian
furrows at right angles to one another in order to subcontinent yet exists. The data presented in this
accommodate two crops simultaneously. Although paper permit the first diachronic assessment of the
there is no conclusive evidence for the so-called consequences of dietary change on tooth size and
"granaries" having been used to store grain, it is dental health among prehistoric South Asians.
reasonable to assume that grain was an important Therefore, the significance of this new South Asian
stapleduringthisperiodofprehistory(fordiscussion, data istwo-fold. First,thesedatamaybeused to iden
seeFentress 1984).Harappansitesalso presenta well tifybiologicalresponses to agricultural intensification
developedcopper/bronzetechnology,whichincludes sharedbyHarappansandotherSouthAsians.Second,
a varietyofvesselsand toolsthatmayhavebeenused comparison of this South Asian data with data from
in the treatment and preparation of cultivated food other regions of the world allows detection of any
stuffs(AllchinandAllchin1982;Wheeler1968). unique or distinctive trends in pathology prevalence
Examination of faunal remains by Meadow (1987, sharedbySouthAsianscommensuratewithincreased
1989,Chapter7inthisvolume)andBelcher(Chapter8 relianceupondomesticatedfoodstuffs(Lukacs1991).
in this volume) suggests that the hunting of wild The second area addressed by this paper involves
animals and fishing constituted important aspects of biologicalaffinities.Thequestionsaddressedhereare:
theHarappansubsistencestrategy.Theseexploitation Who were the Harappans? To whom do they share
patterns are attested by the bones ofa wide range of greatest biological affinities, and does the pattern of
wild speciesrecovered from Harappansitesand indi affinities possessed by Harappansindicatebiological
catesa moreintensive use of wild speciesduring the continuity or discontinuity within the Indus Valley
Harappan period thanatany timesincetheneolithic. throughoutprehistory?
These data suggest that, while the Harappan The rise and fall of the Harappan Civilization has
Civilization was no doubt an agriculturally-based beenthe topicofmuchresearchsincethediscoveryof
society, non-agricultural foodstuffs nevertheless Harappa and Mohenjo-daro. SirJohn Marshall, exca
continued to play a vital role in the Harappan vator of Mohenjo-daro, was an early proponent of
economic system. This continued utilization of non indigenousdevelopmentoftheHarappanCivilization
domesticated foodstuffsmaythereforebeafactor that from pre-Harappan cultures within the Indus Valley
callsinto questionearliermodelsofthedevelopment (Marshall 1931).LaterscholarsbelievedtheHarappan
of the Harappan Civilization which presumed an Civilizationtobetheproductof"stimulusdiffusion,"
overwhelming reliance on intensive agricultural oractualmigrationfromtheWest(Gadd1932;Gordon
exploitationfor increasingcentralization,stateforma 1947,1958;Heine-Geldern1956;Mackay1938;Piggott
tion,andintra-aswellasinter-urbaninterdependence. 1950; Wheeler 1968). Numerous correspondences in
Assessmentoftoothsizeanddentalhealthprovides ceramicwares, seals,metal artifacts, beads,and other
insight into subsistence patterns and dietary change. items ofmaterial culturehave beencited as evidence
Recent studies of dental afflictions by age, sex, and ofcontactbetweentheIndusValleyandMesopotamia.
social status among contemporary African hunter Further investigations within the Indus Valley,
gathererand horticulturalgroupshavedemonstrated Afghanistan, and Iran since 1960onceagain haveled
that differences in subsistence activity patterns are many scholars to support a model of indigenous
accompanied by differences in dental pathology development. Ghosh's (1965) argument for ceramic
Biological Adaptations and Affinities of Bronze Age Harappans 139
continuity between "pre-Harappan" times and the and Sassman 1943; Kumar 1971, 1973; Sarkar 1954;
"Mature" Harappan period was followed by the SewellandGuha 1931).
"IndigenousPlace"and "ThreeInfluences"theoriesof This study not only provides new biological
Dales(1965)and Fairservis(1975). Continued investi evidencefrom the Harappan phasecemetery(R37)at
gation ofnew sites and re-examination of previously Harappa, but simultaneously employs several multi
excavated sites in the late 1960s and early 1970s variate statistical techniques based on three different
brought additional support for indigenous develop types of biological variation (cranial metric, cranial
mentoftheHarappanCivilizationfrompre-Harappan non-metric, and dental non-metric). Employment of
cultures within the IndusValley(Allchinand Allchin several statistical techniques offers the advantage of
1968,1982;Jarrige1982;Jarrigeand Lechevallier1979; avoidingconclusionsdrawnfromanyonealone,asall
JarrigeandMeadow1980;Mugha11970,1990). contain their own inherent biases. Similarly, use of
However, excavations in Iran, Afghanistan, and more than one type of biological variation allows a
western Pakistan have indicated to other workers composite picture of affinities to be drawn from the
(Beale 1973;Biscione1983;Kohl 1978,1979;Lamberg entirety of variation available at this time. This
Karlovsky 1972, 1978; Lamberg-Karlovsky and Tosi approachhelpstoalleviatesuchproblemsofbiological
1973;Santoni 1984;Tosi 1979) that prior to the rise of affinityanalysis asdirectional selection and differen
the HarappanCivilization, the IndusValley was part tiallevelsofsexualdimorphism.
ofan "earlyurbaninteractivesphere" centeredonthe
traders of the Iranian Plateau. Participation of the
Materials and Methods
HarappanCivilizationinthis"interactivesphere"may
haveledto considerableextra-IndusValleyinputand
participationin theriseoftheHarappanCivilization. The Harappan phase cemetery at Harappa is
No less vexing than issues surrounding the rise of locatedto thesouthofMound ABand wasdesignated
the Harappan Civilization, are questions concerning as "CemeteryR37" byShastriin 1937-38(seePossehl,
its demise. Tentatively dated to the period between Chapter 2 in this volume) (Figure 11.1). Subsequent
1900 BC (Dales 1973) and 1750 BC (Agrawal 1966; excavations in the same general area by Wheeler
Allchinand Allchin1968),thedemiseoftheHarappan (1947)andMughal(1968)alsoresultedintherecovery
Civilization has been attributed to Aryan invaders of Harappan phase burials. Earlier in the 1920s and
(ChiIde 1957; Gordon 1958; Piggott 1950; Wheeler 1930s, Vats (1940) excavated in the area between
1968),ecologicalchanges(Dales1966;Kennedy1984a; Cemetery R37 and Mound AB and recovered two
Kenoyer1988;Lambrick1967;Misra1984;Raikes1964, different types of burials referred to as Cemetery H
1965),andsimpleprogressivedegeneration(Fairservis Stratum II (lower or "earth" burials) and Stratum I
1975). (upperor"pot/jar" burials).Bothoftheseburialtypes
Archaeological attempts to resolve questions belongto the LateHarappan phaseandarelaterthan
concerning the rise and subsequent fall of the theburialsintheHarappanphasecemetery(Cemetery
Harappan Civilization have been plagued by prob R37).
lems in dating, inaccessibility of key research areas, Excavations by the University of California,
and interpretations of ceramic and other artifactual Berkeley(UCB)Projecthaveresultedintherecoveryof
assemblages. Recent excavations at the Harappan more than 90 individuals from the Harappan phase
phase cemetery at Harappa allow a unique opportu cemetery (Table 11.1: hereafter termed "R37C"). This
nity to examine these questions from the perspective number compares favorably with the 106reported by
of skeletal biology. Unlike ceramic styles, decorative Dutta (1983) from previous excavations in the same
motifs, and metallurgical technology, genetically general area (Table 11.1: hereafter termed "R37A").
controlled featuresofthe teethand skeletoncannotbe Unfortunatelymanyoftheseindividualsareextremely
transferred verbally or inherited by others not in fragmentary. When well-preserved individuals are
actual contact with the reference population. used as the basis of comparison, 33 individuals were
Unfortunately, many previous analyses of skeletal recovered in 1987 and 1988 and 34 individuals were
remains in South Asia have been limited byan overt recovered from previous excavations. Thus the UCB
racial-typologicalapproachtohumanbiologicalvaria excavationshaveledtoadoublingoftheskeletalinfor
tion,byemploymentofparametersthatincorporatean mationavailablefortheHarappanphaseinhabitantsof
unknown amount of environmental variation, or by thesite.
questionableattemptstoforce thesedata intoa model The R37C sample (from the UCB excavations)
of localized continuity (Dutta 1972, 1975, 1983;Guha represents a young to middle-aged adultpopulation,
and Basu1938;Gupta,Duttaand Basu1962;Krogman for few children and juveniles were recovered.
~
E1400 E1600 E1800 E2000 E2200 E2400 E2600 E2800 c~
N2000 N2000
~
""'t
DryBedofRavlRiver ~
"'1::l
"'1::l
~
tT:l
>-c
N1800 () N1800 (")
~
(j
.~....
Q
c'
;:s
V'J
t-I
1.0
N1600 N N1600 00
0'1
I
t t-I
1.0
1.0
o
~
3::
N1400 N1400
;:
.....
E.:
~
Vi'
-(")
~-..
;:s
N1200 N1200
~
~
~
"'1::l
"'1::l
o""'t
U.C.BerkeleyExcavations:1986-1990 ~
(")
Nl000 Operation1:Cemetery _ Nl000 ;::ro
Operation2:MoundAB and Gateway
Operation3:MoundE .- ,
PreviousExcavations : ,.]
L__-,
o 100 200meters
N0800 N0800
I
E1400 E1600 E1800 E2000 E2200 E2400 E2600 E2800
Figure11.1 LocationoftheHarappanphasecemeterywithinthesiteofHarappa.
Biological Adaptations and Affinities of Bronze Age Harappans 141
Similarly, the number of older adults recovered was defined as "secondary" if retained in jaw fragments.
alsolow(Table11.2). Loose teeth not in association with jaw remains are
Dental remains obtained during the UCB excava defined as "isolated." Dental remains from primary
tions were examined by Lukacs. The dental sample burials constitute nearly half of the entire sample
was recovered from two different burial contexts: (48.1%), whileisolatedremainsaccountfor only 15%.
primary and secondary (Table 11.3). Primary context Malesandfemalesareequallyrepresented,butnearly
remains are composed of 16 undisturbed burials. one-fourth of the sample isderived from individuals
Secondary contexts include incomplete skeletal ofunknownsex.
remains displaced from their original context by Analysis of tooth size involves measurement of
erosion or by later intrusive interments. Dental tooth length and tooth breadth. In this report tooth
remains recovered from disturbed contexts are lengthisdefined asthemesiodistal (MD)diameterof
Table11.1.SexDistributionoftheHarappanPhaseCemeterySamples
Sample Males Females Unknown Total
R37A 38 55 13 106
R37C 19 29 42 90
Total 57 84 55 196
Table11.2.AgeDistributionoftheHarappanPhaseR37CSample
Category AgeRange Number
Subadult <16Years 15
Young Adult 17-34Years 35
Middle-AgedAdult 35-55Years 27
OlderAdult >55Years 13
Total 90
Table11.3.HarappanPhaseR37DentalSamplebyBurialContextandbySex
Context
Primary Secondary Isolated Total
Max Man Total Max Man Total Max Man Total Max Man Total
II 25 24 49 12 11 23 12 6 18 49 41 90
12 21 24 45 11 13 24 10 6 16 42 43 85
C 23 28 51 13 13 26 9 4 13 45 45 90
P3 23 26 49 18 19 37 8 5 13 49 SO 99
P4 27 25 52 16 20 36 4 10 14 47 55 102
Ml 20 20 40 23 28 51 5 9 14 48 57 105
M2 22 21 43 18 30 48 6 12 18 46 63 109
M3 13 19 32 11 21 32 3 4 7 27 44 71
Total 174 187 361 122 155 277 57 56 113 353 398 751
Males Females Unknown Total
Max Man Total Max Man Total Max Man Total Max Man Total
II 17 16 33 16 16 32 16 9 25 49 41 90
12 14 16 30 17 18 35 11 9 20 42 43 85
C 16 17 33 18 21 39 11 7 18 45 45 90
P3 16 18 34 18 25 43 15 7 22 49 SO 99
P4 18 20 38 19 23 42 10 12 22 47 55 102
Ml 18 22 40 16 19 35 14 16 30 48 57 105
M2 17 23 40 15 22 37 14 18 32 46 63 109
M3 13 20 33 6 19 25 8 5 13 27 44 71
Total 129 152 281 125 163 288 99 83 182 353 398 751
142 Harappa Excavations 1986-1990: A Multidisciplinary Approach
the tooth taken at its greatest point. Tooth breadth is tooth-trait combinations scored in all dental series
defined as the greatest distance in the buccolingual were considered. If a specific trait was completely
(BL) plane.Thesediametersand theirproduct(crown absentinmore thanhalfoftheseriesexamined,itwas
area), are determined by methods described by eliminated from consideration. While Sjevold (1973)
Wolpoff (1971). All measurements were made by acceptsatraitifitispresentinonlyoneofthesamples
Lukacs with a Helios needle-point dial caliper cali underconsideration,wefeltthatwiththesmallsample
brated to 0.05 mm, and crown diameters were sizesavailableretentionofinfrequentvariableswould
rounded to the nearest 0.1 mm. Assessment of intra unduly magnify the influenceofchanceoccurrenceof
observeraccuracyisbased onrepeated measurement rare traits. These criteria resulted in a reduction of
of the left side of 30 randomly selected plaster casts. dental morphology traitsincludedinthecomparative
Measurementsessions were separated by a period of analysis to ten traitsscoredas 16tooth-traitcombina
ten months. Mean intra-observer difference is -0.06 tions. In most cases, any degree of traitdevelopment
mm. (s.d.= 0.18),afigure similarto thosereportedby wasconsideredapositiveexpression.Theonlyexcep
otherinvestigators(Hemphill 1991;Keiseretal. 1986; tion to this pattern is hypocone development, for
Kolakowski and Bailit1981). which only full expressions were scored as positive
Pathologicalconditionsofthepermanentteethand manifestations.
jawswerealsoassessedbyLukacs.Pathologicalcondi Contingencychi-squarestatisticswerecalculatedto
tions investigated include: abscesses, antemortem determine whether these non-metric dental traits
toothloss(AMTL),calculus,caries,hypoplasia,hyper detect significant heterogeneity in specific trait
cementosis, pulp exposure, and alveolar resorption. frequencies across all seven dental samples. If the
Methods for assessing the degree of expression of numberofsignificantdifferencesexceedsthenumber
these conditions have been described elsewhere of differences expected by chance alone, these traits
(Lukacs 1989; Lukacs et al. 1989). Frequencies of were accepted as providing adequate data for deter
pathological conditions are assessed by individual mining patterns of relative similarity between these
countandbytoothcount. samples.
Twenty morphological features of the permanent To place theR37C samplein regional perspective,
dentalcrownwereassessedfor55tooth-traitcombina affinities between sites were assessed by three
tionsbyLukacsinaccordancewithcriteriasetforthby differentstatistical methods. Traitfrequenciesbased
LukacsandHemphill(1991a).Intra-observervariation on presence/absence dichotomization were arcsine
was assessed by repeated scoring of 35 tooth-trait transformedaccordingto theformula recommended
combinationsina randomsampleof50plasterdental byGreenand Suchey (1976) to stabilizesamplevari
casts. Observation sessions were separated by more ance.
than one year, and the methods used in quantifying In the first method, arcsine transformed trait
observer variation were those recommended by frequencies were used as input for cluster analysis,
Nichol and Turner (1986). For each discrete traitfour anddendrogramswereconstructedineuclidianspace
percentages were calculated, and the mean for these with Ward's minimum variance technique (Ward
four measuresofintra-observererrorare well within 1963). In the second method, mean measureofdiver
thosequotedbyNicholandTurner(1986). gencedistanceswerecalculatedbetweeneachsample
Frequenciesofdentaltraitswerecalculatedforeach pairbymeansoftheformula recommendedbyGreen
grade of expression present in the Harappan sample and Suchey (1976). The standard deviation of these
according to the individual count method of Scott distances were calculated according to the method of
(1973, 1977, 1980), and trait expression was Sjevold(1973),andstandardizeddivergencedistances
dichotomizedintopresence/absenceonly for compar were calculated by dividing the mean measure of
isonwithotherSouthAsiandental series(Table11.4). divergence of a specific comparison by its standard
In contrast to the method of Sofaer et al. (1986), only deviation. These standardized distances are more
Table11.4.DentalNon-MetricSampleSizesandSources
Sample Abbrev. Date Nmax Source
Harappa HAR Harappanphase 33 ThisReport
ChalcolithicMehrgarh MR2 4500B.C. 25 LukacsandHemphill(1991a)
NeolithicMehrgarh MR3 6000B.c. 49 Lukacs(1988)
Inamgaon INM 1600-700B.C. 41 Lukacs(1987)
Mahadaha MDH 8000B.c. 11 LukacsandHemphill(1991b)
Timargarha TMG 1400-850B.c. 21 Lukacs(1983)
SaraiKhola SKH 200-100B.c. 25 Lukacs(1983)
Biological Adaptations and Affinities of Bronze Age Harappans 143
appropriate for evaluating and comparing relative comparedwithtenothersamplesfromSouthAsiaand
distancesincaseswhere widelydifferentsamplesizes theNearEast(Table11.5).Onlythosesamplesscored
are involved (Sofaer et al. 1986). Since the pattern of by other workers in accordance with the criteria of
variation of seven different samples from one Berryand Berry(1967) wereaccepted. As withdental
another-involving 21 different standardized non-metric variables, each cranial trait had to be
distances-is difficult to visualize, standardized consideredinevery sampleinordertobeincluded in
distances were used as input for multidimensional thecomparativeanalysis.Thisresultedin theelimina
scaling. Kruskal'sStressFormula One (Kruskal 1964) tion of two traits (foramen ovale incomplete, foramen
was used, and results were ordinated into three spinosum open). In addition, those traits completely
dimensional space. absent in more than half of the samples were elimi
The third method used to examinegroupaffinities nated. This resulted in elimination of five traits
betweendentalsamplesisprincipalcomponentsanal (bregmatic ossicle, auditory exostoses, bifaceted
ysis. Arcsine transformed trait frequencies were occipital condyle, palatine torus, maxillary torus).
standardized to have a zero mean and unit variance Finally, traits which may be the product ofmuscular
priorto submission to principalcomponentsanalysis. development (highest nuchal line) or whose positive
Unrotated principle components were used since manifestations are subject to widespread differing
varimax rotation (Kaiser 1958) served to reduce the interpretations(supraorbital foramen complete) were
percentageofthe total varianceexplainedby thefirst eliminated from consideration.Thesecriteriaresulted
three principal components and offered no improve in a reduction in the number of cranial non-metric
ment in interpreting the patterning of component traitsconsideredto18traits.Chi-squarestatisticswere
loadings. Factorscorecoefficients(eigenvectorcoeffi calculated to detect significant differences in indi
cients) for each variable were multiplied by the vidual trait frequencies across all samples. As
standardized arcsine transformed frequency for each describedfornon-metricdental traits,ifthenumberof
sample. These values were summed for each sample significant differences exceeds the number of differ
according to each of the first three principal compo ences expected by chance, these non-metric cranial
nents. The resulting scores were plotted into three variables were considered viable data for examining
dimensions to illustratethepositionofeachsamplein patterns of inter-group variation. Affinities between
multicomponentspace. cranial series was accomplished with the same three
Analysis of cranial non-metric traits was methodsdescribed fordental non-metric traits.
performed in a manner similar to that described for All adultcrania derived from the Harappan phase
non-metric traits of the permanent teeth. Twenty cemetery were assessed for 30 metric variables by
seven non-metric traits were scored for adult Kennedy. All measurements have been standardized
individuals in the R37C sample by Hemphill by both the Biometrika school and by Martin and
according to the criteria of Berry and Berry (1967). Saller (1957). Only those variables for which data are
Because of small sample size, fragmentary condition available for either males or females in each sample
of specimens, and for greater comparability with wereaccepted. This reduced the numberof variables
previously published reports, traits were scored by considered tofourteen. Sampleswere thendividedby
side, and only full expressions oftraits were consid sexandmean valuesobtained.
ered positivemanifestations. Comparativeanalysisofcraniometricvariationisin
In thecomparativephaseofanalysis, thefrequency three parts. In the first part, craniometric variation
of cranial non-metric traits in the R37C sample were encompassedby samplesfrom twoexcavationefforts
Table11.5.CranialNon-MetricSampleSizesandSources
Sample Abbrev. Date Nmax Source
Harappa HAR Harappanphase 24 ThisReport
Egyptian EPT 4~ B.c. 250 Berryand Berry(1967)
AncientPalestinian APAL 700B.C. 54 Berryand Berry(1967)
ModernPalestinian MPAL Modem 18 Berryand Berry(1967)
Punjabi PUN Modem 53 Berryand Berry(1967)
Burmese BUR Modem 51 Berryand Berry(1967)
Mahadaha MDH 8000B.c. 11 ThisReport
Lidar LDR 23~2ooo B.c. 25 KlugandWittwer-Bakofen(1985)
Kamidel-Loz KEL 500B.C. 47 KlugandWittwer-Bakofen(1985)
SaraiKhola SKH 200-100B.C. 26 KlugandWittwer-Bakofen(1985)
Bedouin BED Modem 35 Henkeand Disi (1981)
144 Harappa Excavations 1986-1990: A Multidisciplinary Approach
in the Harappan phase cemetery (R37A, R37C) and secondary inhumations in pots/jars. To control for
from two different burial contexts in the Late different sex ratios presented by each sample, group
Harappan cemetery ("HlJAR" = Stratum I-upper values were calculated as the average between male
jar/potburials, "H20PEN" = StratumII-lowerearth and female means for each craniometric variable.
burials) are compared against five other prehistoric Groupmeanvaluesweresubmittedtoclusteranalysis,
skeletal series from Pakistan. This phase of analysis anddendrogramswereconstructedineuclideanspace
seekstodeterminethelevelofwithin-sitevariabilityat with Ward's minimum variance technique. Principal
Harapparelativetothelevelofvariationencompassed componentsanalysis wasused to provide a checkon
by other prehistoric samples from Pakistan. Samples thepatternsofaffinityderivedfromclusteranalysisas
were divided by sex, and analysis of variance was well as to gain additional insight into the pattern of
used to determine whether these craniometric vari variation among these modem and prehistoricSouth
ables detect significant differences across these six Asiansamples.
samples. As withcranialand dental non-metrictraits, In the final phase, South Asiancraniometric varia
if the number of significant differences exceeds the tion is placed in regional perspective by contrasting
numberofdifferencesexpectedbychance,these vari SouthAsiancranial seriesagainstprehistoric samples
ables were accepted as providing adequate data for from Egypt, Anatolia, Mesopotamia, and the Iranian
determining patterns of relative similarity among Plateau (Table 11.6). To provide maximum compar
theseprehistoricgroups. ability, the same variables used to examine
Two methods were used to assess relative affinity within-South Asiancranial variation wereemployed.
bysexamongprehistoricgroupsfromPakistan.Inthe Group values were derived in the same manner
firstmethod,sex-specificmeanparametervalueswere described for South Asian samples, and these values
submitted to clusteranalysis,anddendrogramswere were submitted to cluster analysis and principal
constructed for males and for females in euclidean components analysis for assessment of biological
spacewithWards'minimumvariancetechnique.Only affinities.
those variables for which data were available in all
samples were accepted. This criteria resulted in a Results
reductioninthenumberofvariablesconsideredtoten
Odontometrics
for females and eleven for males. In the second tech
nique,groupsex-specificmeanparametervaluesused Mean mesiodistal and buccolingual crown dia
inclusteranalysisweresubmittedtoprincipalcompo meters for the sex-pooled sample obtained from R37
nentsanalysis.Thesevalueswerestandardizedbysex are presented in Table 11.7. A paired-samples t-test
acrossgroups to havea zero meanand unitvariance. wasused totestforsignificantdifferencesintoothsize
Unrotated principle components were derived, and between left-right tooth pairs. Only two crown dia
factor score coefficients (eigenvector coefficients) for meters (LI2MD, LM3MD) exhibit significant
each variable were multiplied by the standardized differencesbetweenrightandleftantimeres.Sincethis
mean parameter value for each sample. These values number of significant differences falls below the
were summed for each sample by sex according to number of significant differences expected from
each of the first three principal components, and chance(p<0.05),mean valuespresentedinTable11.7
resultingscoreswereordinatedinto threedimensions reflectmeasurementsfrom theleftsideonly.
to illustrate the position ofeach sample in multicom Mean crown areas (in mm2) by side are also
ponentspace. presented inTable11.7.Meansfor rightand leftsides
The second phaseofanalysisexaminesthelevel of are nearly equal, confirming the lack of asymmetry
craniometric variation among living and prehistoric foundbypaired-samplest-testsofindividualparame
SouthAsians. SamplesofmodemVeddahs,Nepalese, ters.Summedcrownareasbyjawrevealthatmaxillary
and Tibetans were compared with the six prehistoric teeth (631 mm2) are slightr larger than their
samplesfrom Pakistanused inthefirst phaseofanal mandibular isomeres (567 mm ). When mean crown
ysis.SamplesfromtheHarappanphasecemeterywere areas for both right (1189 mm2) and left (1198 mm2)
pooled, but the two samples from Late Harappan areconsidered,individualsfrom theHarappanphase
phaseCemeteryH wereconsidered separately.These cemeterypossessanoverall toothsizeof1194mm2.
latter samples were considered separatelyin order to Total crown area is used to place the tooth size of
determine if differences in inhumation practices are the inhabitants of the Harappan phase cemetery in
accompaniedbydifferencesinbiologicalaffinities.The perspective with regard to other South Asian dental
earlier burials of the Late Harappan cemetery are series. Differencesin totalcrownareaarepresentedin
extendedburialswith potteryarrangedat thehead or two ways: as a histogram of total crown area (Figure
feet. In theupperlevelsofthecemeterytheburialsare 11.2),andasabivariateplotoftotal tooth sizeagainst