Skull Shape Asymmetry and the SocioeconomicStructure of an Early Medieval Central European Society

Lucie Bigoni,1* Vaclav Krajıcek,1,2 Vladimır Sladek,1 Petr Velemınsky,3 and Jana Velemınska1

1Department of Anthropology and Human Genetics, Faculty of Science, Charles University in Prague,12800 Prague 2, Czech Republic2Department of Software and Computer Science Education, Faculty of Mathematics and Physics,Charles University in Prague, 12116 Prague 2, Czech Republic3Department of Anthropology, National Museum, 11579 Prague 1, Czech Republic

KEY WORDS craniofacial asymmetry; socioeconomic structure; Mikulcice castle and sub-castle;geometric morphometrics

ABSTRACT The socioeconomic structure of an earlymedieval society from the Mikulcice settlement (CzechRepublic) was studied on the basis of an evaluation ofthe fluctuating and directional asymmetry (DA) of skulls.Two distinct inhabited regions, castle and sub-castle,were compared. Fluctuating asymmetry (FA) was usedas a bioindicator of environmental stress, which isthought to have been different in the Mikulcice castleand sub-castle regions. DA is consistent with biomechan-ical loading, and it was expected to reflect different sub-sistence patterns. The material consisted of 129 craniafrom what was presumed to be a higher socioeconomicclass (Mikulcice castle) and 71 crania from middle andlower socioeconomic classes (Mikulcice sub-castle). As acomparative sample, 138 crania from modern, lower soci-oeconomic classes (Pachner Collection) were used. The

three-dimensional coordinates of 68 landmarks weredigitized and analyzed using geometric morphometrics.In terms of DA, the highest values were recorded in thesub-castle sample and confirmed their lower socioeco-nomic position, with a grittier and lower protein dietcompared with the castle sample. In terms of FA, distinc-tive differences between the sexes were found. In males,no differences were observed between castle and sub-cas-tle, and the lowest FA values were recorded. In females,significantly higher values of FA were found, surpris-ingly in the castle sample, comparable with the morestressed Pachner Collection. We suspect that the FAreflects a more varied population of castle females as aconsequence of patrilocality, although environmentalstress remains a possibility. Am J Phys Anthropol150:349–364, 2013. VVC 2013 Wiley Periodicals, Inc.

At the beginning of the 9th century AD, the westernbranch of the Slavs managed to establish a state in cen-tral Europe, the so-called Great Moravia. This firstSlavic state (Poulık, 1975; Havlık, 1978), dated to the9th–10th centuries AD, stood politically and culturallyon the frontier of Byzantium, Rome, and the Frankishempire (Polacek, 2008). According to written, archeologi-cal and anthropological sources (Stloukal and Vyhnanek,1976; Havlık, 1978; Poulık, 1985; Velemınsky and Pola-cek, 2008), Great Moravia represents a society with com-plex social stratification. One of the most important cen-ters of Great Moravia was the Mikulcice settlement.Mikulcice was an important communication crossroads(Kvet, 1999), and simultaneously provided favorable liv-ing conditions for its local inhabitants (Polacek, 2008).Since the first investigation in the field in the 1950s(Poulık, 1975, 1985; Klanica, 1985), distinct inhabitedregions have been discovered in the Mikulcice settle-ment: a fortified acropolis (castle), partially fortified sub-castle, and rural hinterland (Polacek, 2008). Archeolo-gists have described social differences between the resi-dential center, with a number of churches, burialgrounds, and a handicraft industry, and its rural hinter-land. However, a question remains regarding how dis-tinctive the socioeconomic differences were between cas-tle and sub-castle (Polacek, 2008; Machacek, 2010).The Mikulcice settlement was located in the valley of

the Morava River. The center, on the sand dunes,enjoyed the natural protection of river channels (Fig. 1).The castle region, with a princely residence and a num-

ber of churches, represented the main residential andspiritual area of the agglomeration (Polacek, 2008). Sev-eral burial grounds have been excavated close to the cen-tral churches, inside the first fortification. The centralburial grounds are considered to represent the highestsocial class of society, as the church and especially relicsare known to have played an important role in the EarlyMiddle Ages (Ohler, 2001; Buko, 2008). Evidently this sep-aration was not strict, because people who were directlyconnected with the princely palace, such as craftsmen andinhabitants providing services for the residence, couldalso be buried in the castle locale (Polacek, 2008).Outside the fortification, other burial grounds existed

in the sub-castle region. It is assumed that representa-tives of the middle and lower socioeconomic classes were

Grant sponsor: The Ministry of Education, Youth and Sports of theCzech Republic; Grant number: MSM 0021620843; Grant sponsor:Project GACR; Grant number: 206/07/0699; Grant sponsor: Ministryof Culture of the Czech Republic (DKRV, 00023272).

*Correspondence to: Lucie Bigoni, Department of Anthropologyand Human Genetics, Faculty of Science, Charles University, Vin-icna 7, 12800 Praha 2, Czech Republic. E-mail: lucfia@volny.cz

Received 31 March 2012; accepted 13 November 2012

DOI 10.1002/ajpa.22210Published online 3 January 2013 in Wiley Online Library

(wileyonlinelibrary.com).

VVC 2013 WILEY PERIODICALS, INC.

AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 150:349–364 (2013)

buried in these burial sites, e.g., Kostelisko, around theIXth church, and Klasterisko (Stloukal and Vyhnanek,1976; Havlık, 1978; Polacek, 2000, 2008). The churchesin the sub-castle are thought to have been parts of theestate founded by nobles in the vicinity of the princelyresidence, which is also a possible explanation for thefinds of rich warrior graves, especially in the Kosteliskoarea (Polacek, 2008). The sub-castle inhabitants evi-dently participated in ensuring the running of the centerand in providing services for the privileged classes,including the military personnel (Klanica, 1987; Stana,1997; Polacek and Marek, 2005; Hladık et al., 2008).The identification of socioeconomic differences in

Mikulcice presents several difficulties. The Mikulcice set-tlement cannot be unequivocally identified as one of the

locations cited in contemporary written sources (Polacek,2008). Written records, such as codes of law, chronicles,patrimonial letters, and military reports, imply signifi-cant proprietary and social differentiation in Moraviansociety (Havlık, 1978, 1993). Archeological findings ofgrave goods, a possible indicator of the socioeconomicstructure of the Mikulcice settlement, could be biased byChristian ritual. On the other hand, Christianity wasfirst adopted by the ruling layer (e.g., Trestık, 2001;Polacek, 2008); the spread of Christianity in the lowersocial classes was not so rapid (Ohler, 2001; Smetanka,2004), and the influence of Christian ritual on funeralpractices could only be found among all social classes aslate as the second half of the 9th century (Trestık, 2001;Polacek, 2008). It is not known how other indicators,

Fig. 1. A map of the Mikulcice settlement [modified according to Polacek (2000)].

350 L. BIGONI ET AL.

American Journal of Physical Anthropology

such as burial location and grave construction, couldclearly differentiate the social class of an individual(Stloukal and Vyhnanek, 1976; Polacek, 2000, 2008).The study of human skeletal remains could possibly be

used for distinguishing different socioeconomic groups(Larsen, 1999; Sladek et al., 2006). However, similar tothe case of written records and archeological research,this biological source also fails to yield a definite resolu-tion regarding the Mikulcice population. No significantdifferences have been found between the castle andother regions when analyzing stature (Dobisıkova et al.,2008), and linear enamel hypoplasia related to a general-ized growth disturbance (Trefny and Velemınsky, 2008).However, significant differences between the expectedhigher and lower classes were noted when analyzing theincidence of caries (Stranska et al., 2008), prevalence ofenthesopathies (Havelkova et al., 2011), and stable car-bon (C) and nitrogen (N) isotopes from bone collagen(Smrcka et al., 2008). These studies indicated a higherbiomechanical stress arising from the diet of the lowersocial classes, compared with the favored millet and agreater volume of meat in the diet of the higher classes.Exceptions to this conclusion were females from the cas-tle, who were more affected by entheseal changes thanexpected (Havelkova et al., 2011).We attempted to answer the question of socioeconomic

differences between castle and sub-castle by using a geo-metric morphometric estimation of craniofacial asymmetry.Our interest was in fluctuating and directional asymmetry(DA), which could reflect the socioeconomic structure of apopulation (e.g., Palmer and Strobeck, 1986; Graham et al.,2003, 2010; Polak, 2003; Van Dongen, 2006; Schaefer andBookstein, 2009). Fluctuating asymmetry (FA) is thoughtto reflect the ability of an organism to cope with geneticand environmental stress during its development (e.g.,Woolf and Markow, 2003; Schaefer et al., 2006; Ozener,2010a; Ozener and Fink, 2010), although there is a largeliterature discussing this capability of FA (e.g., Houle,1998; Gangestad and Thornhill, 1999; Bjorksten et al.,2000; Van Dongen, 2006, 2007; Carcamo et al., 2008).Nevertheless, FA is still used as an indicator of the stressand developmental stability of an organism (Palmer andStrobeck, 1986, 2003; Graham et al., 2010). Therefore, FAis used as a measure of the health, living conditions, andvariability of a population (Palmer, 1994; Doyle and John-ston, 1977; Perzigian, 1977; Kimmerle and Jantz, 2001;Koehler et al., 2004; Willmore et al., 2005; Klingenberg etal., 2010). Some studies have discussed the presumedincreased buffering in the disruption of development infemales relative to males (Klein, 2004; Møller et al., 2009),especially in samples from stressful conditions (Ozener,2010b). Higher levels of FA could also be associated withrapid morphological change within a population (Kimmerleand Jantz, 2002). Craniofacial FA produces different valuesfor different morphological regions (DeLeon, 2007; DeLeonand Richtsmeier, 2009). Geometric morphometric craniofa-cial FA assessment has already been used to estimate dif-ferent levels of environmental stress in humans, e.g.,between people of higher and lower socioeconomic status inTurkey (Ozener, 2010b), individuals from the Early andLate Christian cemetery in Sudanese Nubia (DeLeon,2007), and inbred and outbred samples from an Adriaticpopulation (Schaefer et al., 2006).DA arises from genetic and environmental stress and

it reflects biomechanical loading, which is connectedwith repeatable actions and behavior. Craniofacial DA isprobably most affected by biomechanical pressure

related to masticatory function (Carlson and VanGerven, 1977; Ferrario et al., 1997; Moreira et al., 2008;Gomes et al., 2011); the chewing habit is asymmetricbecause the majority of a population prefers the rightside (Nissan et al., 2004; Diernberger et al., 2008; Marti-nez-Gomis et al., 2009). Masticatory function is con-nected with diet and subsistence adaptation, and thediet of a past population can be used to make interfer-ences about its social stratification (Le Huray and Schut-kowski, 2005). Asymmetric brain growth and, therefore,asymmetric growth of the vault are also associated withthe diet (Menegaz et al., 2010; Cray et al., 2011). Vaultasymmetry occurs in the context of habitual behavior,such as babies’ sleeping positions, cradling, or the ab-sence of unrestricted motility (Manning and Chamber-lain, 1991; Turk et al., 1996; Manning et al., 1997; Cava-lier and Picaud, 2008). Conversely, the cranial base isconsistent in DA (Kwon et al., 2006).The goal of this study was to compare the DA and FA

of early medieval crania found in Mikulcice castle andsub-castle, in order to evaluate socioeconomic differen-ces, using a comparative sample with a low socioeco-nomic status. We expected to find the Mikulcice sub-cas-tle to be intermediate in terms of FA between Mikulcicecastle, with higher socioeconomic groups, and the com-parative sample with the lowest socioeconomic group,because the sub-castle population, according to archaeo-logical finds, represented middle and lower socioeco-nomic groups. In comparing DA, we expected to see adifferent picture. Because Mikulcice persisted on a hard,grittier subsistence diet, and for the comparative samplea nonhard diet has been proposed, we expected that boththe castle and sub-castle in Mikulcice would exhibithigher levels of DA than the comparative sample.

MATERIALS AND METHODS

Cranial samples

A sample of 200 adult individuals of estimated sex(Stloukal and Vyhnanek, 1976; Stloukal, 1989; Velemın-sky et al., 2005) was selected from Mikulcice castle andMikulcice sub-castle (Tables 1 and 2). The collection of theMikulcice population is housed in the Department of An-thropology, National Museum, Prague, Czech Republic.The Mikulcice castle sample, from burial grounds aroundthe IInd, IIIrd, IVth, and XIth churches, included a totalof 129 crania (61 females and 68 males). The Mikulcicesub-castle sample, from the Kostelisko and Klasteriskoburial sites, and around the IXth church, consisted of 71crania (38 females and 33 males).As a comparative sample, we used 138 adult crania of

known sex (65 females and 73 males) from the PachnerCollection, housed in the Department of Anthropologyand Human Genetics, Faculty of Science, Charles Uni-versity, Prague. This collection originates from the 1930sand represents poor people from Prague (Borovansky,1936; Pachner, 1937). These people had been subjectedto high levels of environmental stress (Pachner, 1937),which is corroborated by findings from the sample, suchas skeletal asymmetry (Kujanova et al., 2008), low sex-ual dimorphism (Borovansky, 1936; Pachner, 1937), gra-cility of skeletons (Bigoni et al., 2005), and teeth loss(Stranska et al., 2005).Basic descriptive statistics for some linear dimensions

and the centroid size of the crania from Mikulcice castle,Mikulcice sub-castle, and the Pachner Collection are pre-sented in Tables 1 and 2.

351SKULL SHAPE ASYMMETRY OF EARLY MEDIEVAL SOCIETY

American Journal of Physical Anthropology

Data acquisition

For craniofacial asymmetry assessment, 68 ectocranialanatomic landmarks (10 unpaired and 29 paired) were cho-sen (Martin and Saller, 1957) to provide a high level of ana-tomic detail of the cranium in which the landmarks wereas evenly distributed as possible (Fig. 2; for detail, see Big-oni et al., 2010). On the basis of the expected different lev-els of DA and FA in different parts of the cranium, the land-marks were divided into three regions for study: the upperface (29 landmarks), vault region (24 landmarks), and cra-nial base (23 landmarks). No landmarks around the alveo-lar processes were used for analysis, because of the fre-quent intravital loss of teeth in the Pachner Collection. Alllandmarks were recorded as three-dimensional (3D) coordi-nates, using a MicroScribe G2X contact digitizer (Immer-sion Corp., San Jose, CA). Each cranium, fixed in modelingclay, was digitized in two positions (DeLeon, 2007). Thefirst position, on the cranial base, enabled the recording ofalmost all landmarks chosen in the region of the face andvault. The remaining landmarks, especially on the cranialbase, were then acquired from a position on the vault. In

both positions, three reference points [bregma (B), nasion(N), and lambda (L)] were marked on each cranium as theorigin and x- and y-axis directions, respectively. Thesepoints were used to align all of the landmarks of each cra-nium within a common coordinate system. The combina-tion of the superior and inferior aspects provided a com-plete configuration.

Shape asymmetry analysis

Shape asymmetry can be imagined as the distance ofthe actual shape from its closest symmetrical variant.For the shape asymmetry analysis of the cranium, wecaptured individual asymmetry (IA) for specimen i anddecomposed it into a sum of FA, DA and an error of mea-surement (e) using the method of Mardia et al. (2000)and Schaefer et al. (2006):

IAi ¼ DA þ FAi þ ei

At a general level, the method is based on Bookstein’s(1991) geometric morphometric approach working with

TABLE 1. Basic statistics of craniofacial linear distances and centroid size in femalesa

MIKca F (1) MIKsc F (2) PACH F (3) SignificantdifferencesN Mean 6 SD (mm) N Mean 6 SD (mm) N Mean 6 SD (mm)

Nasion-Lambda 61 171.7 6 6.1 38 168.8 6 4.4 64 164.6 6 5.5 1–3***; 2–3*Euryon-Euryon 57 138.4 6 5.1 38 136.2 6 5.1 64 143.3 6 5.8 1–3**; 2–3***

Bregma-Basion 56 130.1 6 4.9 36 128.9 6 5.4 64 123.3 6 4.6 1–3***; 2–3***

Prosthion-Nasion 53 66.7 6 4.8 31 64.8 6 3.5 49 61.7 6 5.1 1–3***

Zygomax-Zygomax 56 91.8 6 4.8 33 90.2 6 4.2 64 87.1 6 5.0 1–3***

Prosthion-Opisthion 47 126.9 6 6.4 28 126.4 6 5.1 49 121.1 6 7.2 1–3*Nasion-Basion 56 97.0 6 4.5 36 96.1 6 4.1 64 93.4 6 3.7 1–3**

Mastoid-Mastoid 41 102.7 6 4.8 27 100.4 6 4.8 64 99.3 6 4.7

N Centroid Size 6 SD N Centroid Size 6 SD N Centroid size 6 SD

Upper face 27 22.9 6 0.3 13 22.6 6 0.3 46 22.6 6 0.4 1–3*; 2–3*Vault 26 20.9 6 0.2 17 20.8 6 0.3 39 20.7 6 0.3Cranial base 24 20.2 6 0.3 14 20.0 6 0.3 47 19.8 6 0.3

a MIKca, Mikulcice castle; MIKsc, Mikulcice sub-castle; PACH, Pachner’s collection; F, female; M, male; number in parenthesesindicates subgroups.* P\ 0.05; **P\ 0.01; ***P\ 0.001.

TABLE 2. Basic statistics of craniofacial linear distances and centroid size in malesa

MIKca M (4) MIKsc M (5) PACH M (6) SignificantdifferencesN Mean 6 SD (mm) N Mean 6 SD (mm) N Mean 6 SD (mm)

Nasion-Lambda 68 176.7 6 6.1 33 174.4 6 5.7 73 172.3 6 6.3 4–6*Euryon-Euryon 67 142.3 6 5.5 32 139.8 6 4.6 73 147.1 6 6.5 4–6**; 5–6***

Bregma-Basion 66 135.1 6 4.9 32 136.7 6 4.5 73 130.1 6 5.8 4–6***; 5–6***

Prosthion-Nasion 59 70.1 6 4.1 25 67.2 6 5.7 57 66.8 6 4.7Zygomax-Zygomax 65 94.2 6 3.4 26 95.6 6 4.9 71 90.1 6 5.7 4–6**; 5–6***

Prosthion-Opisthion 57 133.3 6 5.6 23 134.0 6 6.9 57 126.9 6 7.7 4–6**; 5–6***

Nasion-Basion 66 102.4 6 4.7 32 102.5 6 4.3 73 99.0 6 4.7 4–6*; 5–6**

Mastoid-Mastoid 52 107.0 6 4.0 23 105.8 6 4.0 73 105.7 6 5.1

N Centroid Size 6 SD N Centroid Size 6 SD N Centroid Size 6 SD

Upper face 34 23.3 6 0.4 10 23.1 6 0.2 52 23.0 6 0.4Vault 28 21.2 6 0.3 13 21.0 6 0.1 41 21.1 6 0.3Cranial base 40 20.6 6 0.3 13 20.7 6 0.3 49 20.4 6 0.5

a MIKca, Mikulcice castle; MIKsc, Mikulcice sub-castle; PACH, Pachner’s collection; F, female; M, male; number in indicatessubgroups.* P\ 0.05; **P\ 0.01; ***P\ 0.001.

352 L. BIGONI ET AL.

American Journal of Physical Anthropology

landmark coordinates. This is a relatively new method ofasymmetry analysis that has the potential to be more ro-bust for assessing developmental instability of complexobjects such as the human skull (Palmer and Strobeck,2003). As a first step, we removed differences in orienta-tion, size, and position by aligning the whole sampleusing generalized Procrustes analysis (Bookstein, 1991),and calculated the IA. Individual shape asymmetry wascomputed by subtracting the reflected relabeled speci-men, aligned back to the original specimen by ordinaryProcrustes analysis, from the original landmark coordi-nates. In this way, one avoids having to define themedial axis explicitly. Directions from reflected pairedlandmarks to the original counterparts can be inter-preted as the directions that the landmark had to movefrom an ideally symmetric position. If unpaired land-marks are included in the analysis, associated directions

express their deviation from an ideal position on themedial axis. Additionally, the average of specimen land-marks and relabeled reflected images form an ideally sym-metric shape that is used as a reference for visualization.The sample was aligned by generalized Procrustes analy-sis, removing the effect of difference in scale. Further, IAwas analyzed using linear discriminant analysis withleave-one-out cross-validation, to evaluate the separabil-ity of the sample with respect to a grouping variable (sexand population). To assess the statistical significance ofthe difference between such groups, we performed Hotel-ling’s T2 test with permutation.As a second step, DA was computed as the average of

IA. DA represents the bilateral variation in a sample,where one side is consistently different from the other inconformation or size (Van Valen, 1962; Palmer and Stro-beck, 1986; Palmer, 1994; Klingenberg et al., 2002). The

Fig. 2. The location of landmarks on the upper face, vault, and cranial base. For better illustration, the landmarks are similarlylinked in Figures 2–4. 1, Bregma; 2, nasion; 3, lambda; 4, nasospinale; 5, prosthion; 6, opisthion; 7, basion; 8, staphylion; 9, stau-rion; 10, foramen incisivum; 11, maxillonasofrontale; 12, maxillofrontale; 13, supraconchion; 14, frontomalare orbitale; 15, ectocon-chion; 16, subconchion; 17, apertion; 18, frontotemporale; 19, frontomalare temporale; 20, jugale; 21, zygotemporale superior; 22,zygotemporale inferior; 23, zygomaxillare; 24, coronale; 25, euryon; 26, sphenion; 27, crotaphion; 28, infratemporale; 29, auriculare;30, entomion; 31, asterion; 32, mastoidale; 33, postalverion; 34, occipitocondylion posterior; 35, occipitocondylion anterior; 36, fora-minolaterale; 37, ovale mediale; 38, basostyloidion anterior; and 39, caroticum mediale.

353SKULL SHAPE ASYMMETRY OF EARLY MEDIEVAL SOCIETY

American Journal of Physical Anthropology

statistical significance of DA can be measured as a frac-tion of DA and FA, summarized using the Frobeniusnorm D2ðX;YÞ ¼ X � Yk k2F and divided by correspondingdegrees of freedom (Kent and Mardia, 2001) in a sampleof N specimens:

T2 ¼ ND2ðX;YÞ=dPN

i¼1 D2ðIAi;DAÞ=ðN � 1Þd

� Fd;ðN�1Þd

Unfortunately, this measure has an F-distribution onlyin the case of isotropic data. This condition is rarely ful-filled in real-world situations. Therefore, we used theProcrustes Anova, implemented in MorphoJ [describedby Klingenberg et al. (2002)], which can handle noniso-tropic data by using full multivariate analysis (MANOVA).Finally, we calculated FA as the difference between

the IA and DA for each individual. Thus, FA representssmall random differences between corresponding partson the left and right side of an individual with DAremoved, i.e., it is interpreted with respect to the popula-tion sample as a population feature (Van Valen, 1962;Palmer, 1994). The vectors of FA were summarized andthe resulting scalar values were described statisticallyand tested by a standard analysis of variance (ANOVA).FA and DA as assessed using landmark-based method-

ology are different from traditional methods of asymme-try analysis (Palmer and Strobeck, 1986) but measurethe same biological phenomena as length measurementdoes in traditional morphometrics (Palmer and Strobeck,2003; Schaefer et al., 2006; Graham et al., 2010). How-ever, we cannot expect absolute numbers with respect toparticular landmarks, as asymmetry is a measure of dis-tance between complete landmark configurations (Pro-crustes distance). Moreover, the differences expressingthe deviation from symmetry depend on the selection ofa landmark set, because Procrustes analysis distributesthe error equally among the landmarks. This behaviordoes not prevent the use of FA and DA for comparison,because shape variability relative to proportions of asym-metry is preserved. The effect of different sizes does notplay as important a role as in traditional morphometry,because asymmetry is considered to be a property of nor-malized objects with respect to a sample. The methodologyused (Palmer and Strobeck, 2003; Schaefer et al., 2006)also allowed us to quantify the FA of the whole set of land-marks without needing to remove landmarks with statisti-cally significant DA when evaluating FA, as recommendedin Palmer and Strobeck (1986) for traditional morphome-try, because in landmark-based methodology DA and FAare both components of the IA of complete landmark config-uration (Mardia et al., 2000; Graham et al., 2010).As measurement error could be a confounding factor

when assessing FA (Palmer and Strobeck, 1986; Robin-son et al., 2002), we digitized all the landmarks of sixrandomly selected specimens six times, with a minimumof 1 day allowed between digitization. We used Pro-crustes ANOVA (Klingenberg and McIntyre, 1998; Klingen-berg et al., 2002, 2010) to quantify measurement error.The results of measurement error assessment are pre-sented in Table 3 and show that the variance as a resultof repeated measurement was significantly smaller incomparison with the FA variance. Measurement erroraccounts for 0.4% of total phenotypic variation (Grahamet al., 2010).Additionally, centroid size (Bookstein, 1991) was used

as a basic characteristic describing size differences

between groups. Centroid size is a unit less measure andcan only be used for comparison of samples described bythe same set of landmarks.Asymmetry evaluation was performed by customized

calculations, according to the description in Mardia et al.(2000) and Schaefer et al. (2006) and implemented in anOctave (version 3.4.2) environment. Furthermore, a cus-tomized application was developed for the purposes ofvisualization and data exploration, using OpenGL andPython (version 2.7.1). Both linear discriminant analysisand statistical tests were computed in MorphoJ (version1.03a). Supplementary computations, tables, and plotswere done in Excel 2003, PAST (version 2.10, Hammeret al., 2001), and R (version 2.13.0).

RESULTS

The descriptive characteristics of the basic craniofaciallinear dimensions, centroid size, and ANOVA results arepresented in Tables 1 and 2. For females, the largest val-ues for linear measurements and centroid size werefound in Mikulcice castle, intermediate values in Mikul-cice sub-castle, and the lowest values in the PachnerCollection. For males, for half of the linear dimensions,Mikulcice castle exhibited greater values than Mikulcicesub-castle, whereas the lowest values were found in thePachner Collection. For both sexes, the Mikulcice sam-ples did not differ significantly from each other. In con-trast, the linear dimensions from the Pachner Collectiondiffered significantly from the Mikulcice samples, espe-cially from Mikulcice castle, but there were no signifi-cant differences in centroid size.The DA shape evaluation is presented in Table 4. All

samples exhibited significant DA in all three morphologi-cal regions. For both sexes, Mikulcice sub-castle gave thehighest values of DA in all morphological regions, espe-cially in the upper face. Intermediate DA values, closerto Mikulcice sub-castle, came from Mikulcice castle, andthe lowest from the Pachner Collection. The highest dif-ferences between samples were noted in the upper facialregion, with the lowest and the most irregular being thecranial base.The DA comparison between samples is presented in

Table 5. For both sexes, Mikulcice castle and Mikulcicesub-castle did not differ in DA, but the Pachner Collec-tion differed significantly from the Mikulcice samples.For females, a somewhat higher degree of DA in Mikul-cice sub-castle was indicated by greater differences

TABLE 3. Procrustes ANOVA for craniofacial shape on the basisof the position of all 68 landmarks

EffectSums ofsquares

Meansquarea

Degrees offreedom F

Individualb 0.1245 0.247 505 6.85***

Sidec 0.0034 0.036 93 1.01Ind 3 Sided 0.0167 0.036 465 27.11***

Errore 0.0077 0.001 5,820

a Mean square was multiplied by 1,000.b Individual effect represents the variation between individualsin the symmetric component of shape.c Side, systematic difference between the original and mirroredcopy of each individual, directional asymmetry.d Ind 3 Side quantifies FA, effects of Side and Ind 3 Side repre-sent asymmetric component of the shape.e Error, the residual variation due to measurement error,includes both symmetric and asymmetric components.*** P\ 0.001.

354 L. BIGONI ET AL.

American Journal of Physical Anthropology

between the Pachner Collection and Mikulcice sub-cas-tle, compared with lower differences between the Pach-ner Collection and Mikulcice castle.DA changes in the shape of the upper face are dis-

played in Figure 3. The Mikulcice samples (Fig. 3a–d)were similar in terms of DA shape. The midsagittal

points, aperthions, and lateral points of the upper facerotated clockwise. The right side of the upper face nar-rowed relatively in the right–left direction, especially inthe lower part. Likewise, the face arch was relativelyhigher on the right side. The margin of both orbitsrotated counterclockwise. The right orbit was relativelylower and wider, whereas the left one was more rounded.Upper face DA in the Pachner Collection (Fig. 3e,f) wasexpressed much less than in the Mikulcice samples. ThePachner Collection exhibited a reverse tendency in therotation shape of the midsagittal plane and in the shapeof the upper face, but a similar tendency in orbital DA.Shape changes of the vault region as a result of DA

are demonstrated in Figure 4. In this region as well, thecrania from Mikulcice were all similar in DA (Fig. 4a–d).The right half of the vault was relatively wider, espe-cially in the posterior part, and also higher. The inferiorpart of the cranium was relatively wider on the left side.In the Pachner Collection, DA was much less expressedthan in the Mikulcice samples, but it had a similartendency (Fig. 4e,f), especially in females. The PachnerCollection differed from the Mikulcice samples in the

TABLE 4. DA for the craniofacial shape assessed by Procrustes ANOVAa

Sample N Sum of squares Mean squareb Degree of freedom F Summarized DA/nc

Upper face (n 5 27)MIKca (F) 27 0.0054 0.138 39 16.8*** 27

0.0083 0.008 1014MIKca (M) 34 0.0074 0.190 39 26.78*** 30

0.0091 0.007 1287MIKsc (F) 13 0.0030 0.077 39 12.08*** 32

0.0030 0.006 468MIKsc (M) 10 0.0025 0.065 39 7.83*** 35

0.0029 0.008 351PACH (F) 46 0.0023 0.059 39 4.77*** 7

0.0216 0.012 1755PACH (M) 52 0.0023 0.059 39 4.23*** 6

0.0277 0.014 1989Vault (n 5 24)MIKca (F) 26 0.0026 0.084 31 3.28*** 17

0.0230 0.034 667MIKca (M) 28 0.0034 0.108 31 5.46*** 20

0.0351 0.031 1131MIKsc (F) 17 0.0025 0.081 31 4.94*** 25

0.0107 0.028 377MIKsc (M) 13 0.0013 0.042 31 2.14*** 17

0.0122 0.035 348PACH (F) 38 0.0015 0.048 31 1.7** 6

0.0471 0.035 1334PACH (M) 41 0.0014 0.045 31 1.68* 6

0.0526 0.038 1392Cranial base (n 5 23)MIKca (F) 24 0.0018 0.061 29 1.77** 13

0.0198 0.026 775MIKca (M) 40 0.0021 0.071 29 2.28*** 9

0.0166 0.020 837MIKsc (F) 14 0.0018 0.061 29 2.14*** 22

0.0082 0.016 496MIKsc (M) 13 0.0016 0.055 29 1.56* 21

0.0074 0.020 372PACH (F) 47 0.0028 0.075 29 2.12*** 8

0.0334 0.028 1178PACH (M) 49 0.0027 0.093 29 2.45*** 10

0.0335 0.027 1240

The first line for each sample is the effect of side for the particular sample and the second line characterizes residual variation.a See Tables 1 and 2 for abbreviations; n, number of landmarks; N, number of specimens.b Mean square was multiplied by 103.c Values were multiplied by 106.* P\0.05; **P\0.01; ***P\0.001.

TABLE 5. P-values of Hotelling’s T2 test of the DA betweensubgroupsa

MIKsc F PACH F MIKsc M PACH M

Upper faceMIKca F 0.305 \0.001 MIKca M 0.279 \0.001MIKsc F \0.001 MIKsc M \0.001

VaultMIKca F 0.304 0.028 MIKca M 0.589 0.0342MIKsc F 0.002 MIKsc M 0.0649

Cranial baseMIKca F 0.720 0.155 MIKca M 0.194 0.081MIKsc F 0.050 MIKsc M 0.644

a See Tables 1 and 2 for abbreviations.*P\ 0.05; **P\ 0.01; ***P\ 0.001.

355SKULL SHAPE ASYMMETRY OF EARLY MEDIEVAL SOCIETY

American Journal of Physical Anthropology

tendency for an anterior shift of the right forehead, to-gether with a posterior movement of the left posteriorpart of the vault.DA changes in the shape of the cranial base are

visualized in Figure 5. All of the samples studied showeda similar trend: the left basis was relatively wider, espe-cially in the posterior part. In Mikulcice females, thepalate was relatively wider and longer on the left.The highest mean FA values were attained in the cra-

nial base (0.011), with intermediate values in the vault

(0.009), and the lowest in the upper facial region (0.006).No significant differences between samples were foundin the vault and cranial base, so only the graph of thecomparison of FA between samples in the upper facialregion is shown (Fig. 6). Unpaired landmarks wereexcluded from this comparison, because they reflectedthe lowest values of FA (DeLeon, 2007). Mikulcice castlefemales exhibited significantly higher FA compared withother Mikulcice samples; their FA values were similar tothe highly stressed Pachner Collection (ANOVA: Mikulcice

Fig. 3. DA of the upper face. Frontal view. Thin lines represent a symmetric mean shape; bold lines indicate the directional asym-metric changes of each landmark. For better visibility, the observed asymmetry has been exaggerated threefold. The black arrowsshow shifts in DA for the 40% of landmarks exhibiting the greatest magnitude. A: Mikulcice castle females; (B) Mikulcice castle males;(C) Mikulcice sub-castle females; (D) Mikulcice sub-castle males; (E) Pachner Collection females; (F) Pachner Collection males.

356 L. BIGONI ET AL.

American Journal of Physical Anthropology

Fig. 4. DA of the vault. Superior view. Thin lines represent a symmetric mean shape; bold lines indicate the directional asymmetricchanges of each landmark. For better visibility, the observed asymmetry has been exaggerated threefold. The black arrows show shiftsin DA for the 40% of landmarks exhibiting the greatest magnitude. A: Mikulcice castle females; (B) Mikulcice castle males; (C) Mikul-cice sub-castle females; (D) Mikulcice sub-castle males; (E) Pachner Collection females; (F) Pachner Collection males.

357SKULL SHAPE ASYMMETRY OF EARLY MEDIEVAL SOCIETY

American Journal of Physical Anthropology

Fig. 5. DA of the cranial base. Inferior view. Thin lines represent a symmetric mean shape; bold lines indicate the directionalasymmetric changes of each landmark. For better visibility, the observed asymmetry has been exaggerated threefold. The blackarrows show shifts in DA for the 40% of landmarks exhibiting the greatest magnitude. A: Mikulcice castle females; (B) Mikulcicecastle males; (C) Mikulcice sub-castle females; (D) Mikulcice sub-castle males; (E) Pachner Collection females; (F) Pachner Collec-tion males.

358 L. BIGONI ET AL.

American Journal of Physical Anthropology

castle males, ***P \ 0.001; Mikulcice sub-castle females,***P \ 0.000; Mikulcice sub-castle males, ***P \ 0.001).Mikulcice sub-castle females exhibited significantlylower FA than the Pachner females (***P \ 0.001). Formales, no differences were found between Mikulcice cas-tle and sub-castle; they both attained significantly lowerFA than Pachner males (Mikulcice castle males, **P \0.002; Mikulcice sub-castle males, **P\ 0.002).

DISCUSSION

We believe that, in our samples, DA reflects biome-chanical loading and FA environmental stress and vari-ability, so craniofacial asymmetry could reflect socioeco-nomic differences between the samples. Differences incraniofacial asymmetry were more distinctive in femalescompared with males. Regarding FA, significant differen-ces were found between Mikulcice castle and sub-castlefor females, such that, surprisingly, Mikulcice castlefemales were found to be more similar to the highlystressed comparative Pachner Collection; males did notdiffer in any way. Regarding DA, significant differenceswere not found between Mikulcice castle and sub-castle,but between the Mikulcice samples and the Pachner Col-lection. The highest DA levels were attained for theupper facial region, with intermediate levels from the

vault, and the lowest and the least distinctive levelsfrom the cranial base.Our results from the FA assessment indicate that

Mikulcice inhabitants, even from the assumed lowersocial classes, lived in favorable conditions comparedwith the Pachner Collection. Good living conditions andweather at that time have already been noted in previ-ous studies of Great Moravia (e.g., Bodri and Cermak,1995; Polacek, 2008). In an FA comparison of medievaland modern skulls, similar to our results, Gawlikowskaet al. (2007) noted higher values of FA in the modernsample. The authors ascribed this to the greater influ-ence of developmental stress and a weaker ability forcompensation (Gawlikowska-Sroka, 2006; Gawlikowskaet al., 2007). For our sample, we hypothesize that FAreflects a greater environmental stress during develop-ment for the lowest socioeconomic group, because thePachner Collection is characterized by poor people fromPrague (Pachner, 1937) who did not have sufficientresources to be buried. Similarly, all other skeletal char-acteristics have indicated that these individuals hadoccupied the lowest socioeconomic positions (Borovansky,1936; Pachner, 1937; Stranska et al., 2005; Kujanova etal., 2008). When comparing Mikulcice with the PachnerCollection, it is also worth noting that the Mikulcice pop-ulation was a compact society and the graves came from

Fig. 6. Box-plots of summarized fluctuating asymmetries for the paired landmarks of the upper faces in six monitored samples(see Tables 1 and 2 for abbreviations).

359SKULL SHAPE ASYMMETRY OF EARLY MEDIEVAL SOCIETY

American Journal of Physical Anthropology

a relatively long period, lasting a century, whereas theindividuals from the Pachner Collection were collectedover 3 years and could represent a highly variable sam-ple.For all morphological regions, the Mikulcice samples

exhibited significantly higher levels of DA than thePachner Collection, with the greatest being in the upperface. The greatest differences between samples in DAwere also found in the upper facial region, because thisis the morphological part of the cranium that respondsmost to functional adaptation (Vig and Hewitt, 1975;Shah and Joshi, 1978; Rossi et al., 2003). We believethat the results of our DA evaluation reflect differentdiets between the Mikulcice castle, sub-castle and Pach-ner Collection samples, as a consequence of different bio-mechanical stresses affecting, in particular, the upperface during mastication (Vig and Hewitt, 1975). TheMikulcice inhabitants, mainly from the sub-castle, hadto support a greater level of biomechanical loading thanthe individuals from the Pachner Collection. Our resultsprobably reflect the solid and gritty diet of Mikulcicesubsistence compared with the Pachner Collection. Thediet of Mikulcice inhabitants included cereals as well aspulses and various nuts and fruits; the teeth abrasionalso indicates a gritty diet that was hard to chew(Stranska et al., 2008; Machacek, 2010). In contrast, thePachner Collection represents poor people with anassumed deficiency in their diet (Pachner, 1937), asreflected by other skeletal indicators [e.g., gracility ofskeletons (Bigoni et al., 2005) and teeth loss (Stranskaet al., 2005)]. We expect that the lowest values of DA areconnected with a less gritty diet and nutritional insuffi-ciency.Similarly, the DA direction of maxilla and cheek bones

was similar in the Mikulcice samples, but had a reversedirection in the Pachner Collection. In the Mikulcicesamples, it indicated a more depressed right side of theface, probably in connection with chewing side prefer-ence, which has been reported in general to be on theright side (Nissan et al., 2004; Diernberger et al., 2008;Martinez-Gomis et al., 2009). We suppose that the direc-tion of DA indicates the pressure during masticationaffecting the upper face. Some previous studies havenoted a dominant left side of the face (Woo, 1931; Vigand Hewitt, 1975; McIntyre and Mossey, 2002; Ercan etal., 2008), whereas others have described right-side dom-inance (Farkas and Cheung, 1981; Ferrario et al., 1994,1995; Shaner et al., 2000). In our study, the left side ofthe face was wider but lower than the right side.The lowest DA differences and, at the same time, the

lowest values for DA came from the cranial base. Earlierstudies have also shown that the cranial base is consist-ent in DA, even when differences in facial asymmetryhave existed (Kwon et al., 2006). Cradleboarding doesnot affect the morphology of the cranial base (Kohn etal., 1995). It is thus possible that divergent living condi-tions and even everyday habitual activities do not leavetraces in this morphological region.With regard to direction, the DA of the orbits, vault,

and cranial base was similar in the Mikulcice samplesand the Pachner Collection, but in the maxilla and cheekbone regions, DA exhibited the reverse direction. Wesuppose that, in the cranial base and vault, the DAdirection reflects asymmetrical brain growth, whichdevelops prenatally (Annett, 2003; Sun and Walsh,2006). It has been reported that, analogous to ourresults, the right hemisphere is usually larger in the

upper and posterior part and smaller in the basicranialregion (Woo, 1931; Hershkovitz et al., 1992; Rossi et al.,2003), and this phenomenon is stable over populations,similar to handedness. Vault DA is caused, according toour results, above all by the moving of the midsagittalplane. Orbits probably rotated following the vault, buttheir lower border responded to a loading similar to themaxilla and cheek bones (Ercan et al., 2008).Regarding any differences between Mikulcice castle

and sub-castle, the differences were more distinctivebetween females than males. We believe that our resultsindicate to some extent different socioeconomic positionsbetween males and females. By using a DA comparison,the expected biomechanical loading to support the upperfacial region was found to be higher in the Mikulcicesub-castle than Mikulcice castle, especially in females.The comparison of cranial size has already shown higherdifferences in Mikulcice females than males. Mikulcicecastle females exhibit consistently larger linear dimen-sions of the skull than Mikulcice sub-castle females. Thissuggests differing diets in the Mikulcice samples, whichwill reflect socioeconomic differences (Walker and Hew-lett, 1990; Le Huray and Schutkowski, 2005). Given thata low-protein diet leads to a smaller cranial size (Cohenand Armelagos, 1984; Sardi et al., 2006), we assumeMikulcice castle females lived in better conditions withbetter subsistence. Similarly, the grittier diet of lowersocioeconomic groups, with the related intense mastica-tory forces, changes the skeletal craniofacial shape moreand results in more expressed DA (Greene et al., 1967;Vig and Hewitt, 1975; Carlson and Van Gerven, 1977;Gomes et al., 2011). A stable C and N isotope analysis ofrich and poor graves from Mikulcice has indicated thathigher socioeconomic groups had a diet with a greaterintake of the much-favored millet and more meat com-pared with the lower social groups (Smrcka et al., 2008).Divergent diets in higher and lower socioeconomicclasses have also been supported by a study of cariesincidence, which was more evident in poorer graves(Stranska et al., 2008; Esclassan et al., 2009; Garcin etal., 2010). The authors ascribed this to a greater volumeof cereals in the diet of lower socioeconomic groups.Regarding the stature of higher and lower socioeconomicgroups in Mikulcice, no significant differences werefound; however, the differences were more distinctive infemales (Dobisıkova et al., 2008). Females were moreaffected by caries than males (Stranska et al., 2008).The most inconsistent differences among Mikulcice

females and males were found with the FA comparison.Whereas no differences were seen between Mikulcicecastle and sub-castle for males, females differed signifi-cantly, with higher FA values in Mikulcice castle. Mikul-cice castle females thus showed greatest similarity to thelowest socioeconomic groups of a modern population(Pachner, 1937). This result was surprising, as Mikulcicecastle females were expected to have been derived fromthe highest groups of society (Stloukal and Vyhnanek,1976; Polacek, 2008) and so to have been less affected byenvironmental stress (Ozener, 2010b). As Havelkova etal. (2011) have noted, Mikulcice castle females weremore affected by entheseal changes than the assumedlower socioeconomic groups. It could be that the femalesburied in Mikulcice castle were derived mostly fromlower socioeconomic groups, such as servants and slaves(e.g., Galuska, 2004). FA could also increase with rapidmorphological change within a population (Kimmerleand Jantz, 2002). So another explanation could stem

360 L. BIGONI ET AL.

American Journal of Physical Anthropology

from the fact that patrilocality and patrilinearity wereconventional practices in early medieval populations(Ennen, 2001; Le Goff and Schmitt, 2008). Females werebrought into a population, and this was most conspicu-ous in higher socioeconomic groups. Politically strategicmarriages are likely to have occurred, with females mov-ing between different groups and different regions. It isvery likely, therefore, that many females buried inMikulcice castle came to Great Moravia from ‘‘abroad’’(Bartonkova et al., 1971; Andelova, 2007). More exten-sive morphological variability, representing differencesin females of higher social status as a result of contactwith foreign countries (e.g., exogamic marriages), hasalso been indicated in previous studies (Hanakova andStloukal, 1966; Stloukal and Vyhnanek, 1976).As no differences in FA were seen between Mikulcice

castle and sub-castle males, it is possible that there wereno socioeconomic differences between the castle and sub-castle regions regarding males. It is possible that manymembers of the middle and lower society, such as war-riors, craftsmen, and servants, were buried in the castleregion side by side with higher classes of the population.Another possibility is that, in the sub-castle burialgrounds, many wealthy people from the vicinity of theresidential palace were buried (Polacek, 2008). Ourresults are supported by observations made of other me-dieval societies, where the incidence of enamel hypopla-sia, Harris lines, and teeth caries indicates the lowersocioeconomic position of women, compared with men,depending on differing access to food resources (Slaus,2000; Esclassan et al., 2009; Lucas et al., 2010).

CONCLUSIONS

1. The craniofacial asymmetry assessment was consist-ent with the expected socioeconomic differencesbetween Mikulcice castle and sub-castle inhabitants.The differences were more obvious among femalesthan among males and to some extent suggest a dif-ferent socioeconomic status.

2. Our DA and FA assessments indicates that the Mikul-cice inhabitants, even from what were assumed to belower social classes, lived in favorable conditions,with lower environmental stress, and subsisted on amore solid diet, which was harder to chew, comparedwith the Pachner Collection individuals.

3. The dietary differences between the socioeconomicgroups in Mikulcice were illustrated by the DA of theupper face and vault. The greatest differences in DAbetween samples were found in this region. Mikulcicesub-castle had to support insignificant but higher lev-els of biomechanical stress, affecting the upper faceand vault, in connection with grittier and more solidfoods, such as cereals, compared with Mikulcice cas-tle, where there was more meat in the diet. Con-versely, the comparative highly stressed sample fromthe Pachner Collection exhibited by far the lowest val-ues of DA in the upper facial region, probably in con-nection with an inadequate diet.

4. Differences in asymmetry were depicted in the shapeof the vault region and also cranial size when sex wasconsidered. Whereas males exhibited a greater simi-larity in the Mikulcice castle and sub-castle regions,greater differences existed in females, such thatMikulcice sub-castle females were smaller and at thesame time more directionally asymmetric in the vaultthan Mikulcice castle females. These results lead us

to believe that females buried in the Mikulcice sub-castle probably subsisted on a grittier and lower pro-tein diet than Mikulcice castle females. These resultsare also supported by the hypothesis that males bur-ied in Mikulcice castle and sub-castle were from moremixed social classes.

5. FA could reflect environmental stress and simultane-ously an increase in the variability of a population. Formales, no differences in FAwere seen between Mikulcicecastle and sub-castle, whereas for females the differen-ces were significant, with the Mikulcice castle femalestending to have higher values. They exhibited FA valuessimilar to the highly stressed Pachner Collection sam-ple. We expect that the higher variability among Mikul-cice castle females compared with the other Mikulcicesamples could be connected with the phenomena of pat-rilocality in early medieval societies. At the same time,the higher values of FA could infer lower social posi-tions, such that at Mikulcice castle many highly envi-ronmentally stressed females were buried.

ACKNOWLEDGMENTS

We are grateful to Lumır Polacek from the Institute ofArchaeology of the Academy of Sciences, Brno, for thearchaeological suggestions and background papers.Christopher Ruff, the associate editor, and two anony-mous reviewers provided very helpful comments toimprove the manuscript. We thank International ScienceEditing for the English language editing.

LITERATURE CITED

Andelova S. 2007. Kapitoly ze zivota velkomoravskych zen vkontextu pısemnych a archeologickych pramenu, Master The-sis. Brno: Masarykova Univerzita.

Annett M. 2003. Myths of first cause and asymmetries inhuman evolution. Behav Brain Sci 26:208–209.

Bartonkova H, Haderka K, Havlık LE, Ludvıkovsky J, Vasica J,Vecerka R. 1971. Magnae Moraviae Fontes Historici IV. Brno:Universita J. E. Purkyne.

Bigoni L, Velemınska J, Bruzek J. 2010. Three-dimensional geo-metric morphometric analysis of cranio-facial sexual dimor-phism in a Central European sample of known sex. Homo61:16–32.

Bigoni L, Zaloudkova M, Velemınska J, Velemınsky P, SeichertV. 2005. The occurrence of directional and fluctuating limbasymmetry in a recently identified collection of human bones.J Natl Mus Nat Hist Ser 174:129–147.

Bjorksten T, David P, Pomiankowski A, Fowler K. 2000. Fluctu-ating asymmetry of sexual and nonsexual traits in stalk-eyedflies: a poor indicator of developmental stress and geneticquality. J Evol Biol 13:89–97.

Bodri L, Cermak V. 1995. Climate changes of the last millen-nium inferred from borehole temperatures: results from theCzech Republic—Part I. Global Planet Change 11:111–125.

Bookstein FL. 1991. Morphometric tools for landmark data: ge-ometry and biology. Cambridge: Cambridge University Press.

Borovansky L. 1936. Pohlavnı rozdıly na lebce cloveka. Praha:Ceska akademie ved a umenı.

Buko A. 2008. The archaeology of early medieval Poland: discov-eries, hypotheses, interpretations. Leiden: Brill.

Carcamo HA, Floate KD, Lee BL, Beres BL, Clarke FR. 2008.Developmental instability in a stem-mining sawfly: can fluctu-ating asymmetry detect plant host stress in a model system?Oecologia 156:505–513.

Carlson DS, Van Gerven DP. 1977. Masticatory function andpost-Pleistocene evolution in Nubia. Am J Phys Anthropol46:495–506.

361SKULL SHAPE ASYMMETRY OF EARLY MEDIEVAL SOCIETY

American Journal of Physical Anthropology

Cavalier A, Picaud J-C. 2008. Prevention de la plagiocephalieposturale. Arch Pediatr Suppl 15:S20–S23.

Cohen MN, Armelagos GJ. 1984. Paleopathology at the originsof agriculture. Orlando: Academic Press.

Cray J Jr, Mooney MP, Siegel MI. 2011. Cranial suture biologyof the Aleutian Island inhabitants. Anat Rec 294:676–682.

DeLeon VB. 2007. Fluctuating asymmetry and stress in a medi-eval Nubian population. Am J Phys Anthropol 132:520–534.

DeLeon VB, Richtsmeier JT. 2009. Fluctuating asymmetry anddevelopmental instability in sagittal craniosynostosis. CleftPalate Craniofac J 46:187–196.

Diernberger S, Bernhardt O, Schwahn C, Kordass B. 2008. Self-reported chewing side preference and its associations with oc-clusal, temporomandibular and prosthodontic factors: resultsfrom the population-based Study of Health in Pomerania(SHIP-0). J Oral Rehabil 35:613–620.

Dobisıkova M, Katina S, Velemınsky P. 2008. Stature of the GreatMoravian population in connection with social status. In: Vele-mınsky P, Polacek L, editors. Studien zum Burgwall von Mikul-cice VIII. Brno: Archeologicky ustav AV CR. p77–91.

Doyle WJ, Johnston O. 1977. On the meaning of increased fluc-tuating dental asymmetry: a cross populational study. Am JPhys Anthropol 46:127–134.

Ennen E. 2001. Zeny ve stredoveku. Praha: Argo.Ercan I, Ozdemir ST, Etoz A, Sigirli D, Tubbs RS, Loukas M,

Guney I. 2008. Facial asymmetry in young healthy subjectsevaluated by statistical shape analysis. J Anat 213:663–669.

Esclassan R, Grimoud AM, Ruas MP, Donat R, Sevin A, Astie F,Lucas S, Crubezy E. 2009. Dental caries, tooth wear and dietin an adult medieval (12th–14th century) population fromMediterranean France. Arch Oral Biol 54:287–297.

Farkas LG, Cheung G. 1981. Facial asymmetry in healthyNorth American Caucasians. An anthropometrical study.Angle Orthod 51:70–77.

Ferrario VF, Sforza C, Miani A Jr, Serrao G. 1995. A three-dimensional evaluation of human facial asymmetry. J Anat186:103–110.

Ferrario VF, Sforza C, Miani A Jr, Sigurta D. 1997. Asymmetryof normal mandibular condylar shape. Cells Tissues Organs158:266–273.

Ferrario VF, Sforza C, Poggio CE, Tartaglia G. 1994. Distancefrom symmetry: a three-dimensional evaluation of facialasymmetry. J Oral Maxillofac Surg 52:1126–1132.

Galuska L. 2004. Slovane–doteky predku. Brno: Moravskezemske muzeum.

Gangestad SW, Thornhill R. 1999. Individual differences in de-velopmental precision and fluctuating asymmetry: a modeland its implications. J Evol Biol 12:402–416.

Garcin V, Velemınsky P, Trefny P, Bagousse AA-L, Lefebvre A,Bruzek J. 2010. Dental health and lifestyle in four earlymediaeval juvenile populations: comparisons between urbanand rural individuals, and between coastal and inland settle-ments. Homo 61:421–439.

Gawlikowska A, Szczurowski J, Czerwinski F, Miklaszewska D,Adamiec E, Dzieciołowska E. 2007. The fluctuating asymme-try of medieval and modern human skulls. Homo 58:159–172.

Gawlikowska-Sroka A. 2006. Radiological and anthropometricanalysis of the symmetry and direction of evolution of skullsfrom some historic populations. Ann Acad Med Stetin 52:107–117.

Gomes SGF, Custodio W, Faot F, Cury AADB, Garcia RCMR.2011. Chewing side, bite force symmetry, and occlusal contactarea of subjects with different facial vertical patterns. BrazOral Res 25:446–452.

Graham JH, Emlen JM, Freeman DC. 2003. Nonlinear dynam-ics and developmental instability. In: Polak M, editor. Devel-opmental instability: causes and consequences. New York:Oxford University Press. p35–50.

Graham JH, Raz S, Hel-Or H, Nevo E. 2010. Fluctuating asym-metry: methods, theory, and applications. Symmetry 2:466–540.

Greene DL, Ewing GH, Armelagos GJ. 1967. Dentition of aMesolithic population from Wadi Halfa, Sudan. Am J PhysAnthropol 27:41–56.

Hammer Ø, Harper DAT, Ryan PD. 2001. Paleontological statis-tics software package for education and data analysis. Paleon-tologica Electronica 4:9.

Hanakova H, Stloukal M. 1966. Staroslovanske pohrebiste vJosefove. Praha: Rozpravy Ceskoslovenske akademie ved 76.

Havelkova P, Villotte S, Velemınsky P, Polacek L, Dobisıkova M.2011. Enthesopathies and activity patterns in the Early Medi-eval Great Moravian population: evidence of division oflabour. Int J Osteoarchaeol 21:487–504.

Havlık LE. 1978. Morava v 9.-10. stoletı. K problematice poli-tickeho postavenı, socialnı a vladnı struktury a organizace.Praha: Studie CSAV.

Havlık LE. 1993. Kronika Velke Moravy. Brno: JOTA.Hershkovitz I, Ring B, Kobyliansky E. 1992. Craniofacial asym-

metry in Bedouin adults. Am J Hum Biol 4:83–92.Hladık M, Mazuch M, Polacek L. 2008. Das Suburbium des

Burgwalls von Mikulcice und seine Bedeutung in der Struk-tur des Siedlungskomplexes. In: Bohacova, Polacek L, editors.Burg—Vorburg—Suburbium. Zur Problematik der Nebenar-eale fruhmittelalterlicher Zentren. Internationale Tagungenin Mikulcice VII. Brno: Archeologicky ustav AV CR. p179–212.

Houle D. 1998. High enthusiasm and low R-squared. Evolution52:1872–1876.

Kent JT, Mardia KV. 2001. Shape, Procrustes tangent projec-tions and bilateral symmetry. Biometrika 88:469–485.

Kimmerle E, Jantz R. 2001. An innovative method for estimat-ing fluctuating asymmetry calculated from 3D cranial and fa-cial coordinates. Am J Phys Anthropol Suppl 114:89–90.

Kimmerle E, Jantz R. 2002. Secular change in craniofacialasymmetry studied by geometric morphometry and Procrustesmethods. Am J Phys Anthropol Suppl 117:95.

Klanica Z. 1985. Nabozenstvı a kult, jejich odraz v archeologick-ych pramenech. In: Poulık J, Chropovsky B, editors. VelkaMorava a pocatky ceskoslovenske statnosti. Praha: Academia.p107–139.

Klanica Z. 1987. Vorbericht uber die Ergebnisse der 32. Gra-bungssaison in Mikulcice (Bez. Hodonın). Prehled vyzkumu1985:35–36.

Klein SL. 2004. Hormonal and immunological mechanismsmediating sex differences in parasite infection. ParasiteImmunol 26:247–264.

Klingenberg CP, Barluenga M, Meyer A. 2002. Shape analysisof symmetric structures: quantifying variation among individ-uals and asymmetry. Evolution 56:1909–1920.

Klingenberg CP, McIntyre GS. 1998. Geometric morphometricsof developmental instability: analyzing patterns of fluctuatingasymmetry with Procrustes methods. Evolution 52:1363–1375.

Klingenberg CP, Wetherill L, Rogers J, Moore E, Ward R, Autti-Ramo I, Fagerlund A, Jacobson SW, Robinson LK, HoymeHE, Mattson SN, Li TK, Riley EP, Foroud T. 2010. Prenatalalcohol exposure alters the patterns of facial asymmetry.Alcohol 44:649–657.

Koehler N, Simmons LW, Rhodes G, Peters M. 2004. The rela-tionship between sexual dimorphism in human faces and fluc-tuating asymmetry. Proc Biol Sci 271:S233–S236.

Kohn LA, Leigh SR, Cheverud JM. 1995. Asymmetric vaultmodification in Hopi crania. Am J Phys Anthropol 98:173–195.

Kujanova M, Bigoni L, Velemınska J, Velemınsky P. 2008. Limbbones asymmetry and stress in medieval and recent popula-tions of Central Europe. Int J Osteoarchaeol 18:476–491.

Kvet R. 1999. Alte Wege im Marchtal von dem Engpaß beiNapajedla bis zum Zusammenfluß mit der Thaya. In: PolacekL, Dvorska J, editors. Probleme der mitteleuropaischen Den-drochronologie und naturwissenschaftliche Beitrage zurTalaue der March. Internationale Tagungen in Mikulcice V.Brno: Archeologicky ustav AV CR. p223–226.

Kwon TG, Park HS, Ryoo HM, Lee SH. 2006. A comparison ofcraniofacial morphology in patients with and without facialasymmetry: a three-dimensional analysis with computed to-mography. Int J Oral Maxillofac Surg 35:43–48.

Larsen CS. 1999. Bioarchaeology: interpreting behavior fromthe human skeleton. Cambridge: Cambridge University Press.

362 L. BIGONI ET AL.

American Journal of Physical Anthropology

Le Goff J, Schmitt JC. 2008. Encyklopedie stredoveku. Praha:Vysehrad.

Le Huray JD, Schutkowski H. 2005. Diet and social status dur-ing the La Tene period in Bohemia: carbon and nitrogen sta-ble isotope analysis of bone collagen from Kutna Hora-Karlovand Radovesice. J Anthropol Archaeol 24:135–147.

Lucas S, Sevin A, Passarius O, Esclassan R, Crubezy E, Gri-moud AM. 2010. Study of dental caries and periapical lesionsin a mediaeval population of the southwest France: differen-ces in visual and radiographic inspections. Homo 61:359–372.

Machacek J. 2010. The rise of medieval towns and states inEast Central Europe: early medieval centres as social and eco-nomic systems. Leiden: Brill.

Manning JT, Chamberlain AT. 1991. Left-side cradling andbrain lateralization. Ethol Sociobiol 12:237–244.

Manning JT, Trivers RL, Thornhill R, Singh D, Denman J, EkloMH, Anderton RH. 1997. Ear asymmetry and left-side cra-dling. Evol Hum Behav 18:327–340.

Mardia KV, Bookstein FL, Moreton IJ. 2000. Statistical assess-ment of bilateral symmetry of shapes. Biometrika 87:285–300.

Martin R, Saller K. 1957. Lehrbuch der Anthropologie. Stutt-gart: Gustav Fischer-Verlag.

Martinez-Gomis J, Lujan-Climent M, Palau S, Bizar J, SalsenchJ, Peraire M. 2009. Relationship between chewing side prefer-ence and handedness and lateral asymmetry of peripheral fac-tors. Arch Oral Biol 54:101–107.

McIntyre GT, Mossey PA. 2002. Asymmetry of the parental cra-niofacial skeleton in orofacial clefting. J Orthod 29:299–305.

Menegaz RA, Sublett SV, Figueroa SD, Hoffman TJ, RavosaMJ, Aldridge K. 2010. Evidence for the influence of diet oncranial form and robusticity. Anat Rec 293:630–641.

Møller AP, Fincher CL, Thornhill R. 2009. Why men haveshorter lives than women: effects of resource availability, in-fectious disease, and senescence. Am J Hum Biol 21:357–364.

Moreira RS, Sgrott EA, Stuker H, Alonso LG, Smith RL. 2008.Palatal asymmetry during development: an anatomical study.Clin Anat 21:398–404.

Nissan J, Gross MD, Shifman A, Tzadok L, Assif D. 2004.Chewing side preference as a type of hemispheric laterality. JOral Rehabil 31:412–416.

Ohler N. 2001. Umıranı a smrt ve stredoveku. Praha: Naklada-telstvı H & H.

Ozener B. 2010a. Fluctuating and directional asymmetry inyoung human males: effect of heavy working condition andsocioeconomic status. Am J Phys Anthropol 143:112–120.

Ozener B. 2010b. Brief communication: facial fluctuating asym-metry as a marker of sex differences of the response to pheno-typic stresses. Am J Phys Anthropol 143:321–324.

Ozener B, Fink B. 2010. Facial symmetry in young girls andboys from a slum and a control area of Ankara, Turkey. EvolHum Behav 31:436–441.

Pachner P. 1937. Pohlavnı rozdıly na lidske panvi. Praha: Ceskaakademie ved a umenı.

Palmer AR. 1994. Fluctuating asymmetry analyses: a primer.In: Markow T, editor. Developmental instability: its originsand evolutionary implications. Dordrecht: Kluwer. p335–364.

Palmer AR, Strobeck C. 1986. Fluctuating asymmetry: measure-ment, analysis, patterns. Annu Rev Ecol Syst 17:391–421.

Palmer AR, Strobeck C. 2003. Fluctuating asymmetry analysesrevisited. In: Polak M, editor. Developmental instability:causes and consequences. New York: Oxford University Press.p279–319.

Perzigian AJ. 1977. Fluctuating dental asymmetry: variationamong skeletal populations. Am J Phys Anthropol 47:81–88.

Polacek L. 2000. Terennı vyzkum v Mikulcicıch: Mikulcice-pru-vodce, svazek 1. Brno: Archeologicky ustav AV CR.

Polacek L. 2008. Great Moravia, the power centre at Mikulciceand the issue of the socio-economic structure. In: VelemınskyP, Polacek L, editors. Studien zum Burgwall von MikulciceVIII. Brno: Archeologicky ustav AV CR. p11–44.

Polacek L, Marek O. 2005. Grundlagen der Topographie desBurgwalls von Mikulcice. Die Grabungsflachen 1954–1992.In: Polacek L, editor. Studien zum Burgwall von MikulciceVII. Brno. p9–358.

Polak M. 2003. Developmental instability: causes and conse-quences. New York: Oxford University Press.

Poulık J. 1975. Mikulcice. Sıdlo a pevnost knızat velkomorav-skych. Praha: Academia.

Poulık J. 1985. Svedectvı vyzkumu a pramenu archeologickycho Velke Morave. In: Poulık J, Chropovsky B, editors. VelkaMorava a pocatky ceskoslovenske statnosti. Praha: Archeolog-icky ustav AV CR. p9–80.

Robinson D, Blackwell P, Stillman E, Brook A. 2002. Impact oflandmark reliability on the planar Procrustes analysis oftooth shape. Arch Oral Biol 47:545–554.

Rossi M, Ribeiro E, Smith R. 2003. Craniofacial asymmetry indevelopment: an anatomical study. Angle Orthod 73:381–385.

Sardi ML, Novellino PS, Pucciarelli HM. 2006. Craniofacialmorphology in the Argentine Center-West: consequences ofthe transition to food production. Am J Phys Anthropol130:333–343.

Shah SM, Joshi MR. 1978. An assessment of asymmetry in thenormal craniofacial complex. Angle Orthod 48:141–148.

Shaner DJ, Peterson AE, Beattie OB, Bamforth JS. 2000.Assessment of soft tissue facial asymmetry in medically nor-mal and syndrome-affected individuals by analysis of land-marks and measurements. Am J Med Genet 93:143–154.

Schaefer K, Bookstein FL. 2009. Does geometric morphometricsserve the needs of plasticity research? J Biosci 34:589–599.

Schaefer K, Lauc T, Mitteroecker P, Gunz P, Bookstein FL.2006. Dental arch asymmetry in an isolated Adriatic commu-nity. Am J Phys Anthropol 129:132–142.

Sladek V, Berner M, Sailer R. 2006. Mobility in Central Euro-pean Late Eneolithic and early Bronze Age: femoral cross-sec-tional geometry. Am J Phys Anthropol 130:320–332.

Slaus M. 2000. Biocultural analysis of sex differences in mor-tality profiles and stress levels in the late medieval popula-tion from Nova Raca, Croatia. Am J Phys Anthropol111:193–209.

Smetanka Z. 2004. Legenda o Ostojovi. Praha: NakladatelstvıLidove noviny.

Smrcka V, Velemınsky P, Buzek F, Zocova J. 2008. Stable C, Nisotopes in human skeletal material from the Great Moravianburial site at Mikulcice-Kostelisko. In: Velemınsky P, PolacekL, editors. Studien zum Burgwall von Mikulcice VIII. Brno:Archeologicky ustav AV CR. p169–175.

Stana C. 1997. Mikulcice a Prazsky hrad. Arch Rozhledy 49:72–85.Stloukal M. 1989. Paleodemografie staroslovanskych populacı

na uzemı Ceskoslovenska, PhD Thesis. Praha: Karlova Uni-verzita.

Stloukal M, Vyhnanek L. 1976. Slovane z velkomoravskychMikulcic. Praha: Academia.

Stranska P, Likovsky J, Velemınska J. 2005. State of dentalhealth and selected pathological conditions in the recentskulls from the Pachner’s collection. Slov Antropol 8:127–136.

Stranska P, Velemınsky P, Velemınska J. 2008. The state of den-tition in the Great Moravian population—a comparison of theMikulcice centre and its hinterland. In: Velemınsky P, PolacekL, editors. Studien zum Burgwall von Mikulcice VIII. Brno:Archeologicky ustav AV CR. p121–139.

Sun T, Walsh CA. 2006. Molecular approaches to brain asymme-try and handedness. Nat Rev Neurosci 7:655–662.

Trefny P, Velemınsky P. 2008. Linear enamel hypoplasia in anearly medieval population of Great Moravia. In: VelemınskyP, Polacek L, editors. Studien zum Burgwall von MikulciceVIII. Brno: Archeologicky ustav AV CR. p141–149.

Trestık D. 2001. Vznik Velke Moravy. Praha: NakladatelstvıLidove noviny.

Turk AE, McCarthy JG, Thorne CH, Wisoff JH. 1996. The ‘‘backto sleep campaign’’ and deformational plagiocephaly: Is therecause for concern? J Craniofac Surg 7:12–18.

Van Dongen S. 2006. Fluctuating asymmetry and developmentalinstability in evolutionary biology: past, present and future. JEvol Biol 19:1727–1743.

Van Dongen S. 2007. What do we know about the heritability ofdevelopmental instability? Answers from a Bayesian model.Evolution 61:1033–1042.

363SKULL SHAPE ASYMMETRY OF EARLY MEDIEVAL SOCIETY

American Journal of Physical Anthropology

Van Valen L. 1962. A study of fluctuating asymmetry. Evolution16:125–142.

Velemınsky P, Likovsky J, Trefny P, Dobisıkova M, VelemınskaJ, Polacek L, Hanakova H. 2005. Großmahrisches Graberfeldauf ‘‘Kostelisko’’ im Suburbium des Mikulcicer Burgwalls.Demographie, Spuren nicht spezifischer Belastung physiologi-schen und physischen Charakters auf Skeletten, Gesundheits-zustand. In: Polacek L, editor. Studien zum Burgwall vonMikulcice VI. Brno: Archeologicky ustav AV CR. p539–633.

Velemınsky P, Polacek L. 2008. Studien zum Burgwall vonMikulcice VIII. Brno: Archeologicky ustav AV CR.

Vig PS, Hewitt AB. 1975. Asymmetry of the human facial skele-ton. Angle Orthod 45:125–129.

Walker PL, Hewlett BS. 1990. Dental health diet and social sta-tus among Central African foragers and farmers. Am Anthro-pol 92:383–398.

Willmore KE, Klingenberg CP, Hallgrımsson B. 2005. The rela-tionship between fluctuating asymmetry and environmentalvariance in rhesus macaque skulls. Evolution 59:898–909.

Woo TL. 1931. On the asymmetry of the human skull. Biome-trika 22:324–352.

Woolf CM, Markow TA. 2003. Genetic models for developmental ho-meostasis: historical perspectives. In: Polak M, editor. Develop-mental instability: causes and consequences. New York: OxfordUniversity Press. p99–115.

364 L. BIGONI ET AL.

American Journal of Physical Anthropology