AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 117:113–121 (2002) Dental Relatedness Corresponding to Mortuary Patterning at Huaca Loro, Peru Robert S. Corruccini* and Izumi Shimada Department of Anthropology, Southern Illinois University, Carbondale, Illinois 62901-4502 KEY WORDS odontology; intracemetery; multivariate; bootstrap ABSTRACT Within and between tombs at the 1,000year-old site of Huaca Loro on north coastal Peru, interment characteristics vary to an intriguing degree. Following and elaborating upon prior intracemetery studies, biological relatedness among associated burial groupings was assessed using 23 dental characters (assuming familial allele segregation) for 29 individuals. Biological patterning was based on multivariate distance between individuals using all traits, rather than the previously widespread reliance upon univariate comparison of each Huaca Loro is a truncated pyramidal mound with elite shaft tombs located in the Poma National Historical Sanctuary, Batán Grande, northern coastal Peru. The tombs pertain to the Middle Sicán culture, circa AD 1000, a period of fluorescence and expansion of a theocratic, hierarchical society. It was ultimately overcome around AD 1400 by the Chimú, a coastal rival of the Incas. The Sicán Archaeological Project has worked in the Batán Grande region continuously for over 23 years (Shimada, 1981, 1990, 1995, 2000). Since 1990, excavations have included recovery of five individuals from the East Tomb (3 ⫻ 3 ⫻ 11 m deep: Shimada, 1995; Shimada and Merkel, 1993), and more recently of 24 burials from the deep and elaborate West Tomb (10 ⫻ 6 ⫻ 15 m deep), as well as five burials from the external North Trench (Shimada et al., 1998, 2000) (see Fig. 1). The tomb excavations relate to Shimada’s longterm effort at social reconstruction (and more recently demographic reconstruction) through multidisciplinary analysis of mortuary contexts and contents, and he continues to collaborate with others in comparative osteological and mitochondrial DNA analyses (Shimada et al., 2001). This will increment previous work describing skeleto-dental aspects of geographically related (Peruvian-Andean) populations, including Stewart (1943), Goaz and Miller (1966), Devoto and Arias (1967), Rothhammer and Silva (1989, 1990, 1992), Nelson (1997), and Verano (1997a–c, 1999). A single, inferred elite male was at the bottom of the West Tomb in its central chamber (Fig. 2), accompanied by much material wealth and adjacent to niches containing inferred sacrificial young women. 2002 WILEY-LISS, INC. DOI 10.1002/ajpa.10020 © trait separately within samples. This multivariate approach did seem more informative. Statistically significant variation of biological similarities and dissimilarities corresponded to spatial groupings and also to various specific archaeological indications of the cohesiveness, or lack thereof, of interment pattern. The partition of biological distances among tombs at Huaca Loro supports the archaeological evidence that the tombs represent a planned elite cemetery. Am J Phys Anthropol 117:113–121, 2002. © 2002 Wiley-Liss, Inc. Three meters higher in an antechamber, a juvenile, possibly male, who seemed to be arranged to “look at” the principal interment, plus 18 females (two opposing groups of nine; Farnum et al., 1998) were found in a condition suggesting that these were remains brought in and redeposited from burial elsewhere (Shimada et al., 1998). Many of the undisturbed burials on this antechamber floor were missing extremities such as phalanges, and were accompanied by broken and incomplete ceramic vessels. Some skeletons were disarticulated to a degree difficult to reconcile with any natural postdepositional shifting. These lines of evidence suggest that many of the antechamber females were reburials or already mummified rather than sacrificed at the time of interment of the principal personage. Earlier, the East Tomb had yielded a central male deposited with 1.2 tons of diverse grave goods, a juvenile and a child atop this, and two adult females placed in positions suggesting a symbolic birth-giving scene (Shimada, 1995; Shimada et al., 1998). A variety of questions arise which biology might better answer than archaeology, regarding the relationships among the individuals in these tombs. Were the West Tomb elite male and the overlying Grant sponsor: Shibusawa Ethnological Foundation; Grant sponsor: Southern Illinois University-Carbondale. *Correspondence to: R.S. Corruccini, Department of Anthropology, Southern Illinois University, Carbondale, IL 62901-4502. E-mail: RCORRUCC@SIU.EDU Received 15 February 2000; accepted 30 August 2001. 114 R.S. CORRUCCINI AND I. SHIMADA Fig. 1. Location of Huaca Loro and its excavation areas. nonsacrificial remains relatives, implying a sort of family mausoleum? Were any of these biological kin with the external (north trench) or the East Tomb individuals? Do the males evince higher relatedness than females overall, as would be consistent with patrilocality (or would the opposite happen, suggesting matrilocality)? Do all individuals evince subequal similarity, suggesting (depending upon comparison with some outgroup) either random accumulation or that burials represent nuclear family members but excluded collateral relatives? If this comparison can be based on strong markers of biological relatedness, patterns of relatedness can be uncovered, as prior work suggests. Prior studies Lane and Sublett (1972) influenced subsequent studies by demonstrating probable residence-pattern information in Seneca skeletal biological traits. Increased similarity among males compared to females was thought to indicate social change from prior matrilocal toward patrilocal habits (contra Corruccini, 1998). Supporting evidence would show more similarity preserved across generations within males, while females would have had more disparate origins. Ortner and Corruccini (1976), using skeletal and dental measurements and qualitative traits, found greater internal female relatedness in one inferred matrilocal Iroquoian group and the opposite in two other non-Iroquoian groups of prehistoric Virginia Indians. Owsley and Jantz (1978) traced intracemetery patterning in the Arikara. Corruccini et al. (1982) approached the spatial patterning of interments at a burial ground of African slaves in Barbados. In concordance with some accounts of African family burial practices, they found three spatial clusters involving 102 interments to have a significant partition of between-group to within-group variance. Bondioli et al. (1986) based a larger-scale study on the Iron Age necropolis of Alfedena, Italy, which entailed examining the distribution of 196 different biological traits over 89 people found in three areas. Both males and females showed more similarity within areas and more difference between areas than expected at random, but males showed this pattern more significantly. One possible explanation, albeit one that could not be confirmed, was that for patrilocal extended family burial areas, some female consanguineous relatives were also included. Bondioli et al. (1986) reviewed other case studies of nonrandom grouping of skeletal variants by cemetery area, suggesting family burial DENTAL AFFINITIES AT HUACA LORO 115 Fig. 2. Map of West Tomb, which contained most of the burials. plots or genetic segregation between social strata. Rubini (1996) reconfirmed aspects of genetic endogamy within the Alfedena tomb circles, using cranial discrete traits. Konigsberg (1988) modeled and tested genetic patterning of postmarital residence in prehistoric Illinois sites. More recently, Prowse and Lovell (1996) demonstrated probable endogamous kin patterning between high- and low-status predynastic Egyptian cemeteries. Howell and Kintigh (1996) argued for a nonrandom pattern of genetic relatedness across the cemeteries at the Southwestern US site of Hawikku, and then went a step further and argued that elite individuals, i.e., “leaders,” were probably related, which suggests inherited status. Alternate conclusions slightly at variance with Howell and Kintigh (1996) concerning biological and social patterning according to sex were suggested by Corruccini (1998). Earlier, Corruccini (1972) had analyzed three Pueblo Indian skeletal collections from different villages, and had found significantly higher osteometrical levels of similarity among females in two villages known from ethnohistorical accounts to be matrilo- cal (including Hawikku). This indicated a local continuity of female lineages. A third village of unknown residence pattern reversed that trend, and had females that were relatively discordant. Present objectives Dental characteristics are clearly an informative source of data for establishing genetic segregation within lineages and differences between populations (e.g., Kelley and Larson, 1991; Scott and Turner, 1997; Corruccini and Sharma, 1985; Corruccini et al., 1988), comparing very favorably to other skeletal features (Sjøvold, 1984; Falk and Corruccini, 1982). As the Huaca Loro sample preserved many teeth without inordinate attrition problems (owing to the rather narrow range of age-at-death variation, as most individuals were 25–35 years old), we thought a dental comparison according to burial complex location would be illuminating. We undertook this study to look for possible familial clusters or other biological patterning that might correspond to archaeological patterning. 116 R.S. CORRUCCINI AND I. SHIMADA MATERIALS AND METHODS Indications of biological lineage were sought within and between burial groupings. The total sample was subdivided into the following nine groups: the principal West Tomb central burial (1), the juvenile male “looking” at him (2), the two adequately preserved accompanying (possibly sacrificial) females in the south and north niches (3, 5), eight scorable females to the south (4), eight scorable females from the grouping to the north (6), five extraneous inferred “commoners” from the north trench (7), and the principal interment (8) and three other individuals (9) from the East Tomb (see Fig. 2). To partition variation, these comparisons would entail 69 within-group pairwise similarities and 337 between-group interindividual differences, if this were done on the basis of the individual rather than the trait. Dental variables Silicon molds were taken by R. Benfer and I.S. of maxillary and mandibular arches in Peru, and dental stone casts were made by W. Duncan at Southern Illinois University-Carbondale (SIUC). The dental traits yielding dissimilarities were taken largely from Corruccini and Potter (1981) and Turner et al. (1991). The ranked states of the latter were frequently modified by R.S.C., after a pass through the material, to magnify the sometimes-limited variation seen within this group. The best-preserved side was scored. Several traits were scored and measured from the molds rather than casts because there was a clearer delimitation in the negative in molds. The following traits were scored as shown below: 1. Maxillary central incisor labial convexity, scored in five phases spanning states 1–3 of Turner et al. (1991), i.e., as 1, 1⫹, 2, 2⫹, and 3; 2. Maxillary incisor shovel shape, scored in five ranked phases corresponding directly to states 2– 6 of Turner et al. (1991); 3. Maxillary double shoveling, labial marginal ridging graded into five expressions corresponding roughly to intermediate scores between the six states 0 –5 in Turner et al. (1991), i.e., 1⫺, 2⫺, 3⫺, 4⫺, and 5⫺; 4. Mandibular incisor shoveling, in nine ordinal states spanning states 0 –5 of Turner et al. (1991). These divide zero (no expression), 1 (moderate), and 2 (strong) into quarters for scoring purposes; 5. Canine accessory ridge and basal tubercle, an impression gained jointly from maxillary and mandibular canines of lingual cervical swelling, in six ranked states of development corresponding directly to states 0 –5 in Turner et al. (1991); 6. Maxillary distal premolar buccal cusp (paracone) diameter, the most variable of four maxillary premolar crown components. As with trait 7, this was measured directly on the molds (internally) rather than casts (externally) for greater accu- racy in locating the maximum mesiodistal cusp diameter; 7. Mandibular distal premolar lingual component mesiodistal diameter, essentially the metaconid plus any entoconid; 8. Hypocone development on maxillary M1, four stages corresponding only to states 4 –5 in Turner et al. (1991), i.e., 4, 4⫹, 5⫺, and 5; 9. Hypocone development on maxillary M2, four states scored as in Dahlberg’s older 3⫹/31/2/4⫺/4 system, which corresponds to states 3– 4 in Turner et al. (1991); 10. Maxillary M3 metacone, five stages corresponding somewhat to intermediates of states 0 –5 (as with trait 3 above) in Turner et al. (1991); 11–13. Cusp number for mandibular M1–3, with seven gradations (6, 5⫹, 5, 5⫺, 4⫹, 4, and 4⫺) ranging over states 0 – 4 of Turner et al. (1991); 14 –15. Chord from mesial fovea to central fovea for mandibular M1–2; 16 –17. Chord from central fovea to distal fovea (or distal marginal ridge) for mandibular M1–2; 18 –19. Chord from central to distal fovea on maxillary M1–2. All of 14 –19 were measured with needle-pointed Mitutoyo calipers calibrated to 0.02 mm on the molds, because the features were easier to pinpoint as convex rather than concave; 20. M1 bilaterally missing in the presence of M2; 21. M2 bilaterally missing in the presence of M1; 22. Central incisor winging, whether in upper or lower dentition, scored present/absent conforming to states 1 or 3 in Turner et al. (1991); and 23. Third cusp development on the distal mandibular premolar. Analyses Before undertaking tests to determine whether the distribution of individuals in the tomb was nonrandom, we assessed the replicability of the scoring of dental traits. The distribution of individuals in the tomb can be assessed with univariate as well as multivariate data sets, and we chose to do both. The assessment of the distribution of univariate and multivariate measurements involved analyses of variance (ANOVA). These steps in our analyses are described in greater detail below. Replicability of dental measurements. R.S.C. recorded very small chords and Cartesian coordinate-based measures such as traits 6 –7 and 14 –19 on tooth crowns in similar sorts of studies (Corruccini, 1977a–c, 1978, 1988; Corruccini and Potter, 1981), where the information content was much greater than for traditional maximum diameters of teeth, but the relative magnitude of measurement error was higher. Therefore, multivariate replicability of scoring was examined by comparison of first and second passes through the material (in the same order according to new numbers given the speci- DENTAL AFFINITIES AT HUACA LORO mens, but blind to their identity), and discrepancies ⬎0.1 mm or more than one character state were entered. Confidence was assessed by the tendency of the same cases to cluster together in weighted pair group method (WPGM) dendrograms based on multivariate distance. The advantage of multivariate contrast of individuals (across all traits) as opposed to long lists of sample-bound univariate comparisons seems obvious in looking at the cumbersome analyses of Bondioli et al. (1986) and Howell and Kintigh (1996), which have to be subject to large amounts of Bonferroni redundancy error of type I. Multivariate comparison was approachable in the present study due to common preservation and scorability of most of the traits in each individual, especially mandibular teeth. Thus the three varieties of multivariate dissimilarity coefficients below were calculated between each specimen, and patterns were sought according to the sample partitions above. Euclidian distance coefficients were calculated between individuals using three sets of data as noted below. Euclidian distances provide a nonparametric approach to pattern recognition that is appropriate for mixed continuous and multistate ordinal data such as these. The distance is the square root of average squared difference, based on the normalized scores and measures. Therefore it is a sort of average amount of standard deviation units by which two cases’ scores differ. Analyses were based on 1) raw (but normalized) data, 2) the c-scores of Howells (1989) (normalized by row and column), and 3) geometric shape vectors (each individual divided through by the geometric mean) to equalize robusticity (average character state magnitude) prior to normalization. For the last method, a value of 10 was added to each variable, so that small and zero scores did not skew the shape variables. Shape vectors ostensibly lessen sex difference in robusticity, development, and size, which could not be accomplished by ordinary separate-sex or sexnormalized means because males and females were so unevenly distributed over the small samples. Males have larger teeth, and larger teeth are more complex and tend toward more cusps (Garn et al., 1966), tubercles, and ridges (Scott and Turner, 1997). The finding by Harris (1980) of greater female development of incisor ridging seems to be the exception that proves the rule. These linear dental phenotypic distances should reflect genetic distance, but are not necessarily directly proportional to inverse genetic covariance (refer to the many studies cited in Chapter 4 of Scott and Turner, 1997). They will, however, be directly proportional to the square root of the generalized phenotypic distance and to the extent that the latter quantity reflects genetic distance and variance/covariance, so the linear distances should be directly proportional to the square roots thereof. Doi et al. (1986) assert that odontometric similarity is reflective of actual genetic kinship, down to the level of 117 cousins. They also promote the value of multivariate assessment of similarity among individuals. Shinoda et al. (1998) also rely on odontometric proportionality to reflect kinship in Jomon remains, finding that the cases of highest dental similarity share the same mtDNA haplotypes. Analyses of distribution of univariate and multivariate data sets. Standard one-way ANOVA was first applied to all original variables individually, according to the nine-way division of subsamples. The univariate analysis of all 23 traits using ANOVA based on n ⫽ 29 divided into nine partitions is, of course, very awkward because of multiple tests based on very small and unequal samples, but is provided to contrast with the multivariate approach. Also, although ANOVA is reputedly very robust, its use with the ordinal variables (especially those with few states) might be questionable. The interindividual multivariate distances, by contrast, are quite continuous ratio-interval variables, and ANOVA was also applied to those partitioned according to individuals making up the subsamples. The considerable redundancy of total degrees of freedom incurred by the latter approach (growing from the univariate 28, to 58 ⫻ 57/2 ⫺ 1 ⫽ 1,623 pairwise replicate distances) could be corrected clumsily by using an estimated reduction in degrees of freedom. The multivariate statistical significance of variance in distances was better estimated by bootstrap resampling to simulate the standard deviation of the partitioned distances, according to a theoretical distribution representing the null hypothesis (that all are sampled randomly from the observed total distribution). In this situation, the bootstrap solution (e.g., Mooney and Duval, 1993; Bruce et al., 1995) was especially valuable because it can resample empirical distributions of a variable according to how the variable is generated, and answer a question that classic inferential statistics cannot, without considerable theoretical computational effort. The present problem looked at variation both within and between samples according to the pairwise distances of their members (e.g., t-tests of means of within and between group distances), and was interested in nonrandom distributions of both, while ANOVA usually compares between-sample variance to see if it is significantly larger than within samples. Our partition by tomb context created 40 categories of within- and between-group variation: two consisted of 28 pairwise comparisons (8 ⫻ 7/2) within each group of eight females, there were 10 comparisons within the five trench “commoners” (5 ⫻ 4/2), etc., plus 64 distances in the 8 ⫻ 8 comparison between the two large female groups, plus 40 (8 ⫻ 5) between each female group with the five trench individuals, and so on. The total came out to 69 distances in four within-group samplings of distances, and 337 distances in 36 between-group combinations. Did the heterogeneities and similarities 118 R.S. CORRUCCINI AND I. SHIMADA TABLE 1. Average linear dental shape-vector distance within and between components of the Huaca Loro sample 1. West Tomb principal interment 2. Juvenile “attendant” 3. South “sacrifice” 4. South females 5. North “sacrifice” 6. North females 7. North trench 8. East Tomb principal interment 9. Other East Tomb Within 2 3 4 5 6 7 8 9 — 1.47 1.36 1.48 1.65 1.52 1.55 1.27 1.44 — — (1.25) — (1.57) (1.26) — — — — — — — — 1.31 — — — — — — 1.39 1.12 — — — — — 1.33 1.46 1.54 — — — — 1.50 1.33 1.40 1.65 — — — 1.46 1.28 1.29 1.45 1.44 — — 1.42 1.56 1.57 1.66 1.56 1.37 — 1.44 1.23 1.32 1.55 1.52 1.36 1.45 (1.20) — — — — — — — — both between and within the sample subdivisions follow a random distribution? The 1,624 total distances (magnified by a factor of 4 because of sampling each individual twice for the remeasurement phase) were used for extracting 500 independent random resampling bouts of 406 (nonredundant) distances in the 40 groupings described above, and the standard deviations and their confidence limits across groupings were calculated. RESULTS Replicability Measurement error effects are a first item of concern. Looked at on a univariate basis, some of the intraobserver repeatabilities seem unimpressive. For instance, mandibular incisor shoveling shows an average 6% error (with 13% standard deviation), which seems high, but the contribution of this variable to variance patterns might nevertheless be relatively strong. The multivariate cluster analysis assesses intraobserver repeatability on the level of identifying individual specimens. The two time-successive readings on 29 individuals yield 58 individuals to be clustered according to increasing distance (dissimilarity), and 28 of 29 times the first two specimens to form a cluster are the separate readings for the same individual. The one exception involves the juvenile specimen from the East Tomb, which, owing to maxillary breakage and exfoliating teeth, was missing more variables (12 of the 23) than any other case, and the second scoring for that individual joins as the very next step with its correct complement. Therefore, the cluster analysis yields a very high level of confidence in terms of individual identity overcoming measurement error. Another way of looking at the situation is the following: there is a mean distance of 0.18 between pairs of assessments of single individuals, while the overall mean distance between different pairs of individuals is 1.41 (with a minimum observation of 0.71, which was between two occupants of the North Trench). Whatever the percent error is in remeasuring individuals, it pales by comparison with the variability between individuals. Multivariate approaches The sharpest multivariate partition of distance variance among the nine sample subdivisions is according to shape vectors, giving F ⫽ 16.78. The raw data and c-score data produce F ratios 2–3 points lower, and we assume the shape vectors are both the best expression of morphological difference between lineages and the most apt to lessen sex differences. Therefore, consideration henceforward is limited to those results for the sake of simplicity (however, all three methods yielded very highly correlated distances: r ⫽ ⫹0.94). Table 1 gives average Euclidian distances among all the sample partitions in this study. Distribution of univariate and multivariate data In univariate analysis, only 1 trait of 23 reaches a Bonferroni level of significantly low probability (0.05/23 ⫽ 0.0022), which is the maxillary M1 hypocone: it is most strongly developed among the two (north and south) groups of eight females. The univariate approach yields some indication of difference but scant interpretation in this situation. That the multivariate Euclidian distances are partitioned nonrandomly can be tested by ANOVA of the within-group and between-group distances. This yields the F statistic of 16.78, but including both measured and remeasured individuals exaggerates the pairwise distance sample by a factor of 2 ⫻ 2 or 4. Correcting for that would yield F ⫽ 4.19, with 35 and 370 degrees of freedom (P ⬍ 0.0001). Overcorrecting to the minimal univariate availability of 8 and 20 degrees of freedom yields P ⬎ 0.50, however, with an unknown (but substantial) type II error. The average standard deviation for individual distances from random resampling is found to be 0.0903 for the null hypothesis, and the upper 95% confidence limit of that is 0.1158 from bootstrapping. The observed actual standard deviation (s.d.) is 0.1308. The random resampling s.d. reaches the high observed level of variability only P ⫽ 0.006 of the time. A conventional F-ratio of 0.13082/ 0.09032 ⫽ 2.098 is also indicated if we can trust the parametric quality of the distribution, which would DENTAL AFFINITIES AT HUACA LORO 119 Fig. 3. Depiction of odontological shape separation between groupings or individuals, the numbers showing how much the Euclidian distance exceeds the smallest observed interindividual difference (0.71). Only relatively small residual distances of 0.60 or less are shown (by arrows) to indicate major linkages according to similarity. Numbers in parentheses are average residual distances within individuals of the grouping. indicate P ⬍ 0.02. Thus there is significant heterogeneity of distances across and within the sample partitions. DISCUSSION Significant variance among the biological relations of these groupings is indicated. There does seem to be (probably genetic) nonrandom patterning to dental variation over Huaca Loro’s interments. However, interestingly, no interindividual distance between pairs of samples reaches post hoc significance, whether by the F-ratio of within-group variance or the t-test of mean difference of betweengroup distance. The overall pattern is significant, but no individual component confidently registers greater input to that pattern than others. Neverthe- less, the following general tendencies probably contribute the most to overall heterogeneity (see Fig. 3). Among large contributors to the nonrandom pattern of variation is the relatively small intrasample variance within the eight south females and within the five north trench individuals. This could indicate biological homogeneity such as would result from their having been affected by endogamy or from their representing some sort of kin lineages. Other factors contributing to the significant result are relatively high similarity between those two groups above, and larger distances to all other groups. The propinquity of south females to north trench individuals owes largely to shared strong development of maxillary shoveling, canine tubercles, maxillary distal premolar paracone expansion, and large me- 120 R.S. CORRUCCINI AND I. SHIMADA sial to central fovea maxillary M1 chords. A second principal variance factor was the large distance shown by the eight north females from all other groupings. Furthermore, the north females are relatively (and quite consistently) very heterogeneous among themselves as well, which could suggest, for example, that they might be unrelated marital immigrants into a patrilocal context (among possible speculations). The niche “sacrificial” females are consistent with the pattern shown by the adjacent groupings of females. The south niche female is very close to the other eight south females, who in turn are relatively closely related among themselves, while the occupant of the north niche is well-removed from all other individuals, as are also the north females among themselves and to all others. That the West Tomb principal interment and the East Tomb principal interment share a mutually closer relation than with any of the other individuals and groupings surely is “meaningful,” although no statistical significance can be claimed for this single observation. The similarity resides in similar scores for lengthened lingual aspect and extra cuspidation of the mandibular distal premolar, and identical mandibular M1 mesial to central fovea chords. The similarity among the two principal interments might have been greater but for their missing different teeth. Given some of the even lesser distances among south females and north trench occupants, it might be idly speculated that if this represents family relationship, it is more likely a second-degree biological relationship (e.g., uncle-nephew) rather than first-degree (e.g., siblings, parent-offspring). Each distance between pairs of subgroups, however, is not significant by itself in a post hoc statistical context; only the overall pattern is significant. The north/south females’ within-group variance ratio comes closest to being a significant pairwise comparison but is not formally significant at P ⬍ 0.05, even in an a priori context. Results are highly similar in pattern when comparing the other raw and c-score distances with these shape-vector distances. One notable discrepancy arises in the raw distance between the West Tomb principal and the juvenile, which is relatively reduced, initially suggesting to us some special degree of relationship. Failing to see this clearly expressed in other coefficients, and noting the missing variables for the juvenile, we have reservations about emphasizing this point. At any rate, it does not appear that the west juvenile shares any unique or derived mtDNA pattern with the west principal, whereas the East and West Tomb principal interments may share one such haplotype (Shimada et al., 2001). Social relationships suggested by various archaeological data may indicate possible kinship relations among burials. The juvenile in the West Tomb antechamber was positioned to be looking directly at the principal interment, and varied offerings were placed along the line of sight. The young individual in the East Tomb also appeared to be directly looking at that tomb’s principal personage. A social linkage between the antechamber women in the south part of the West Tomb and the principal personage is suggested by long cloth strips that physically linked them. One woven strip descended from the southwest sector of the antechamber floor to the central chamber, and partially wrapped the upper torso of the principal personage. Another strip covered part of the south antechamber floor and extended to and covered part of the central chamber. No comparable connecting cloth strips were found in the northern half of the central chamber, where biological differences increased. Additionally, ceramics and painted textiles affiliated with the earlier Moche style and iconography are concentrated on the north side, while typical Sicán-style ceramics predominate on the south side. CONCLUSIONS This study traces intracemetery biological patterning, using multivariate interindividual distance comparisons. This contrasts with the sample-dependent univariate and multivariate approaches seen in previous intracemetery studies. The partition of biological distances among burials at Huaca Loro independently supports the archaeological evidence that the tombs represent a planned elite cemetery. The spatial distances and social indicators within and between tombs are indicative of probable family relatedness, or lack thereof, and probable inherited elite status for some. The correspondence of spatial and social relations with dentally based biological distances almost certainly represents some sort of biokinship within the inferred social groupings. However, it is impossible to specify the exact nature of those relationships. ACKNOWLEDGMENTS Excavations at Huaca Loro and related laboratory analyses between 1990 –1997 by members of the Sicán Archaeological Project were supported by two grants generously provided to I.S. by the Shibusawa Ethnological Foundation, Tokyo. R.S.C. received a Summer Research Fellowship from Southern Illinois University-Carbondale for this analysis. Dr. Kazuhiro Mine and Julie Farnum are conducting a complementary osteological study of this material. Dr. Ken-ichi Shinoda recently completed his mtDNA analysis of some 50 individuals, including all those discussed in this paper. LITERATURE CITED Bondioli L, Corruccini RS, Macchiarelli R. 1986. Familial segregation in the Iron-Age community of Alfedena, Abruzzo, Italy, based on osteo-dental trait analysis. Am J Phys Anthropol 71:393– 400. Bruce P, Simon J, Oswald T. 1995. Resampling Stats user’s guide. Arlington, VA: Resampling Stats, Inc. DENTAL AFFINITIES AT HUACA LORO Corruccini RS. 1972. The biological relationships of some prehistoric and historic Pueblo populations. Am J Phys Anthropol 37:373–388. Corruccini RS. 1977a. Crown component variation in hominoid lower third molars. Z Morphol Anthropol 68:14 –25. Corruccini RS. 1977b. Cartesian coordinate analysis of the hominoid second lower deciduous molar. J Dent Res 56:699. Corruccini RS. 1977c. Crown component variation in the hominoid lower second premolar. J Dent Res 56:1093–1096. Corruccini RS. 1978. Crown component analysis of the hominoid upper first premolar. Arch Oral Biol 23:491– 494. Corruccini RS. 1988. Morphometric replicability using chords and Cartesian coordinates of the same landmarks. J Zool 215:389 – 394. Corruccini RS. 1998. On cemetery kin groupings at Hawikku. Am Antiq 63:161–163. Corruccini RS, Potter RHY. 1981. Developmental correlates of crown component asymmetry and occlusal discrepancy. Am J Phys Anthropol 55:21–31. Corruccini RS, Sharma K. 1985. Within- and between-zygosity variance in oral traits among US and Punjabi twins. Hum Hered 35:314 –318. Corruccini RS, Handler JS, Mutaw RJ, Lange FW. 1982. Osteology of a slave burial population from Barbados, West Indies. Am J Phys Anthropol 59:443– 459. Corruccini RS, Sharma K, Potter RHY. 1988. Comparative genetic variance of dental size and asymmetry in U.S. and Punjabi twins. Paris: Mem Mus Natl Hist Nat [C] 53:47–53. Devoto FCH, Arias NH. 1967. Shovel-shaped incisors in early Atacama Indians. J Dent Res 46:1478. Doi N, Tanaka Y, Funakoshi K. 1986. A method for estimation of kinship based on the tooth measurements and its application to the ancient human skeletal remains. J Anthropol Soc Nippon 94:147–162. Falk D, Corruccini RS. 1982. Efficacy of cranial versus dental measurements for separating human populations. Am J Phys Anthropol 57:123–127. Farnum J, Corruccini R, Mine K, Shimada I, Vega-Centeno R, Curay V. 1998. Sicán funerary customs: an interdisciplinary perspective. Paper presented at the 26th Annual Midwest Conference on Andean and Amazonian Archaeology, Anthropology and Ethnohistory, University of Illinois at Urbana-Champaign. Garn SM, Lewis AB, Kerewsky RS. 1966. Groove pattern, cusp number, and tooth size. J Dent Res 45:970. Goaz PW, Miller MC. 1966. A preliminary description of the dental morphology of the Peruvian Indian. J Dent Res 45:106 – 119. Harris EF. 1980. Sex differences in lingual marginal ridging on the human maxillary central incisor. Am J Phys Anthropol 52:541–548. Howell TL, Kintigh KW. 1996. Archaeological identification of kin groups using mortuary and biological data: an example from the American Southwest. Am Antiq 61:537–554. Howells WW. 1989. Skull shapes and the map. Pap Peabody Mus 79. Kelley MA, Larsen CS, editors. 1991. Advances in dental anthropology. New York: Wiley-Liss, Inc. Konigsberg LW. 1988. Migration models of prehistoric postmarital residence. Am J Phys Anthropol 77:471– 482. Lane RA, Sublett AJ. 1972. Osteology of social organization: residence pattern. Am Antiq 37:186 –201. Mooney CZ, Duval RD. 1993. Bootstrapping: a nonparametric approach to statistical inference. Newbury Park, NJ: Sage. Nelson GA. 1997. Dental analysis and determination of occupational activities in a Peruvian formative population. Dent Anthropol 12:10 –14. Ortner DJ, Corruccini RS. 1976. The skeletal biology of the Virginia Indians. Am J Phys Anthropol 45:717–722. Owsley DW, Jantz RL. 1978. Intracemetery morphological variation in Arikara crania from the Sully site (39SL4), Sully County, South Dakota. Plains Anthropol 23:139 –147. Prowse TL, Lovell NC. 1996. Concordance of cranial and dental morphological traits and evidence for endogamy in ancient Egypt. Am J Phys Anthropol 101:237–246. 121 Rothhammer F, Silva C. 1989. Peopling of Andean South America. Am J Phys Anthropol 78:403– 410. Rothhammer F, Silva C. 1990. Craniometric variation among South American prehistoric populations: climatic, altitudinal, chronological, and geographic contributions. Am J Phys Anthropol 82:9 –17. Rothhammer F, Silva C. 1992. Gene geography of South America: testing models of population displacement based on archeological evidence. Am J Phys Anthropol 89:441– 446. Rubini M. 1996. Biological homogeneity and familial segregation in the Iron Age population of Alfedena (Abruzzo, Italy), based on cranial discrete traits analysis. Int J Osteoarchaeol 6:454 – 462. Scott GR, Turner CG. 1997. The anthropology of modern human teeth. Cambridge: Cambridge University Press. Shimada I. 1981. The Batán Grande—La Leche Archaeological Project: the first two seasons. J Field Archaeol 8:405– 446. Shimada I. 1990. Cultural continuities and discontinuities on the northern North Coast, Middle-Late Horizons. In: Moseley ME, Cordy-Collins A, editors. The northern dynasties: kingship and statecraft in Chimor. Washington, DC: Dumbarton Oaks. p 297–392. Shimada I. 1995. Cultura Sicán: dios, riqueza y poder en la costa norte del Perú. Lima: Banco Continental. Shimada I. 2000. Late Prehispanic coastal states. In: Minelli LL, editor. Pre-Inka states and the Inka world. Norman: University of Oklahoma. p 49 –110. Shimada I, Merkel J. 1993. A Sicán tomb in Peru. Minerva 4:18 –25. Shimada I, Corruccini R, Farnum J, Mine K, Vega-Centeno R, Curay V. 1998. Sicán population and mortuary practice: a multi-dimensional perspective. Paper presented at the 63rd Annual Meeting of the Society for American Archaeology, Seattle. Shimada I, Gordus A, Griffin JA. 2000. Technology, iconography, and significance of metals: a multi-dimensional analysis of Middle Sicán objects. In: McEwan C, editor. Pre-Columbian gold: technology, iconography, and style. London: British Museum Press. p 28 – 61. Shimada I, Shinoda K, Corruccini R, Farnum J. 2001. Biological and social dimensions of Sicán burials: integrating artifact, DNA, dental, and dietary analyses. Paper presented at the 66th Annual Meeting of the Society for American Archaeology, New Orleans. Shinoda K, Matsumura H, Nishimoto T. 1998. Genetical and morphological analysis on kinship of the Nakazuma Jomon people using mitochondrial DNA and tooth crown measurement. Zooarchaeology 11:1–21. Sjøvold T. 1984. A report on the heritability of some cranial measurements and non-metric traits. In: van Vark GN, Howells WW, editors. Multivariate statistical methods in physical anthropology. Dordrecht: Reidel. p 223–246. Stewart TD. 1943. Skeletal remains with cultural associations from the Chicama, Moche, and Veru valleys, Peru. Proc US Nat Mus 93:153–185. Turner CG, Nichol CR, Scott GR. 1991. Scoring procedures for key morphological traits of the permanent dentition: the Arizona State University dental anthropology system. In: Kelley MA, Larsen CS, editors. Advances in dental anthropology. New York: Wiley-Liss, Inc. p 13–31. Verano J. 1997a. Physical characteristics and skeletal biology of the Moche population at Pacatnamu. In: Donnan CB, Cock GA, editors. The Pacatnamu papers, volume 2: the Moche occupation. Los Angeles: Fowler Museum of Cultural History. p 189 – 214. Verano J. 1997b. Advances in the paleopathology of Andean South America. J World Prehist 11:237–268. Verano J. 1997c. Human skeletal remains from Tomb 1, Sipán (Lambayecque river valley, Peru); and their social implications. Antiquity 71:670 – 682. Verano JW. 1999. Modified human skulls from the urban sector of the pyramids of Moche, northern Peru. Latin Am Antiq 10:59 – 70.