Efficacy of cranial versus dental measurements for separating human populations.код для вставкиСкачать
AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 57~123-127(1982) Efficacy of Cranial Versus Dental Measurements for Separating Human Populations DEAN FALK AND ROBERT CORRUCCINI Department of Anatomy and the Caribbean Primate Research Center, University of Puerto Rico, Medical Sciences Campus, G.P.O. Box 5067, Sun J u a n Puerto Rico 00936 (D.F.), and Department of Anthropology, Southern Illinois University, Carbondale, Illinois 62901 (R.C.) KEY WORDS Craniometry, Race, Odontometry, Analysis of variance ABSTRACT Cranial and dental measurements were taken on 100 skulls, consisting of 20 specimens each, representing five different human populations. In addition to traditional cranial and dental measurements, sets of nonstandard basicranial and cervical dental measurements were collected. The data were subjected to univariate and multivariate analyses to determine what kinds of data best sort human populations. Both univariate and multivariate analyses indicate that traditional and nontraditional cranial measurements separate major human groups better than do dental measurements, with traditional cranial measurements providing the best sorting criteria. The cervical dental measurements proved to be better sorters of human groups than the traditional maximum dental measurements for univariate, but not for multivariate, analyses. Thus, when utilizing univariate techniques, osteologists can increase sample sizes of teeth by collecting cervical data. Traditional cranial measurements have been used to sort human populations for over one hundred years (e.g., Broca, 1874; Hrdlicka, 1907; De Quatrefages and Hamy, 1882). Recently, Howells (1969) addressed the problem of how one determines which cranial measurements best distinguish human populations and suggested that multivariate techniques be applied to identify measurements that are optimal for analyzing human variation. Dental measurements have also traditionally been used to sort human populations (Hanihara, 1977; Harris and Nweeia, 1980; Sofaer, et al., 1971). There are few if any studies of the relative efficacy of cranial versus dental measurements. The purpose of this paper is to conduct such a study using both univariate and multivariate methods of analysis. Maximum mesial-distal and buccal-lingual dental measurements have traditionally been used in racial and paleontological comparisons, although attrition severely affects maximum dental measurements (Wolpoff, 1971). Since tooth wear is severe enough in some groups to destroy discriminative power, cervical measurements taken at the cemento-enamel junction may be preferable to traditional maximum 0002-948318215701-0123$02.00c 1982 ALAN R LISS, INC. measurements for intraspecific (Black, 1978, 1979)and interspecific (Corruccini, 1977)comparisons. Therefore, in addition to comparing cranial with dental measurements, this study tests the relative sorting efficacy of traditional and nontraditional dental measurements. Similarly, although Howells’ (1969) osteometric study shows that certain traditional cranial measurements do distinguish between human populations, he notes that “the base of the skull (has been) generally neglected and not properly covered” (1969:455).A third purpose of the present study is to test the relative merits of some measurements on the inferior base of the skull and traditional cranial measurements for distinguishing human populations. MATERIALS AND METHODS Cranial and dental measurements were taken on one hundred skulls housed at the United States National Museum of Natural History. Five major human populations were each represented by ten skulls from adult feReceivpd April 16. 19R1: accepted September 14. 19R1 124 D. FALK AND R. CORRUCCINI males and ten skulls from adult males. The five groups include representatives from Terry Collection blacks and whites (collectedin the U.S. during the second quarter of the twentieth century); Eskimos from Point Hope, Alaska (1929); Mongolians from Urga (1912); and Amerindians from Jersey County, Illinois (1945). Terry Collection blacks are estimated as admixed 20-30% with whites (Corruccini, 1974; Angel, personal communication). The Eskimo, Mongolian, and Amerindian samples are more homogeneous than the Terry Collection blacks and whites. A cautionary note is that any of these three homogeneous samples could represent breeding isolates and therefore may not be typical of their larger parent racial groups. Four sets of measurements were taken (from the left sides of skulls when unilateral) by one of us (D.F.) over a period of several weeks. Set 1 The six traditional cranial measurements include maximum length, maximum breadth, face height (nasion-prosthion), nasal breadth, bizygomatic breadth and basicranial distance (nasion-basion). Set 2 Unlike the traditional skull measurements, these six small measurements do not span more than one bone (i.e., they represent small, localized traits), some are innovative, and all are located near the basioccipital area: basioccipital length (basion-hormion), basioccipital breadth (minimum breadth of basiocciput), posterior nares maximum breadth, maximum carotid diameter (once thought a correlate of cranial capacity), foramen magnum breadth and posterolateral jugular-anteromedial foramen lacerum distance. These traits are located on the portion of skull that is often preserved relatively intact or undeformed in archeological remains, whereas the entire skull is rarely preserved as an intact unit. Set 3 The six traditional dental measurements include maximum length upper canine, maximum breadth upper canine, maximum length upper third premolar, maximum breadth upper third premolar, maximum length upper first molar, and maximum breadth upper first molar. These particular teeth were selected in order to sample representatives of different dental fields and therefore to maximize re- trieval of independent information. In each case, the first (polar) tooth was selected because it is the relatively least variable and most highly intercorrelated tooth within its field (Dahlberg, 1945; Henderson, 1975). Set 4 Maximum cervical length and breadth measurements were taken at the cemento-enamel junction of each of the polar teeth measured in Set 3 (cervical measurements are figured in Corruccini, 1977). In one group (Eskimos) delayed repeat measurements were taken for the entire sample. Measurement error averaged 1.3% of the smaller reading for craniometrics, and 3.6% for odontometrics. Statistics Basic descriptive statistics and univariate (F-tests)and multivariate (Wilks’lambda) variance ratios were computed from the data (Table I). Six measurements were selected for each of the four sets because this is an appropriate number of measurements for a statistically powerful multivariate analysis in conjunction with sample sizes of n = 20. Wilks’ lambda was used as the criterion for judging intergroup difference in the multivariate analysis. Wilks’ lambda is the ratio of the determinants of the within-group and of the amonggroup plus within-group variance-covariance matrices. The smaller this ratio is, the more effective the separation of the ellipsoids of samples in the multivariate space (Anderson, 1958). Conversely, larger values of the standard univariate variance ratios (F-ratios) indicate more between-group and less withingroup variance. These variance ratios constitute meaningfully comparable criteria of discriminability, as each analysis includes the same number of variables and specimens. RESULTS The mean univariate variance ratio for six traditional cranial measurements is 7.49, that for the set of nontraditional cranial measurements is 5.40 (Table 1).The maximum dental measurements have an average univariate variance ratio of 2.67, while the nontraditional cervical dental measurements yield a mean ratio of 3.28 (seeTable 1).Thus univariate analysis indicates that both traditional and nontraditional cranial measurements separate racial groups better than do dental measurements, with traditional cranial measurements providing the best sorting criteria. 7.34 8.18 6.53 8.89 10.09 11.15 5.59 7.76 4.73 8.20 7.91 10.65 Maximum dental measurements canine length canine breadth P' length P' breadth M' length M' breadth Cervical dental measurements canine length canine breadth P' length P' breadth M ' length M' breadth (0.54) (0.72) (0.39) (0.58) (0.39) (0.64) (0.47) (0.71) (0.77) (0.56) 10.821 (0.57) (2.94) (2.19) (2.18) (0.71) (1.65) (2.16) 'Each human population is represented by 20 skulls. *Not statistically significant at p = 0.05. See Lext tor discussion. 29.77 21.96 28.83 7.18 31.66 35.89 (7.41) (5.98) (4.92) (1.68) (6.50) (6.46) 5.81 8.16 5.21 8.83 8.13 10.95 7.70 8.43 7.39 9.73 10.42 11.48 29.93 22.40 29.60 6.73 28.68 33.77 (0.48) (0.72) (0.44) (0.70) (0.81) (0.64) (0.41) (0.71) (0.48) (0.65) (0.70) (0.64) (3.30) (2.32) (2.73) (1.12) (2.41) (2.11) (7.55) (6.88) (5.28) (2.05) (6.24) (7.87) Blacks 182.15 136.30 69.48 25.69 128.15 99.35 ~~~ ~ 6.00 7.70 4.95 8.32 8.25 11.10 7.40 8.01 6.52 9.11 9.88 11.74 32.75 23.07 30.17 7.35 31.00 35.95 182.20 137.25 71.57 23.42 138.30 105.60 (0.40) (0.55) (0.41) (0.64) (0.45) (0.46) (0.65) (0.52) (0.87) (0.64) (0.78) (0.47) (1.96) (1.42) (1.92) (0.55) (1.82) (1.82) (5.13) (4.58) (4.32) (1.85) (7.00) (4.72) Eskimos.~ (8.21) (4.24) (5.50) (2.07) (7.05) (5.07) 5.69 7.54 4.80 8.29 7.72 10.66 7.46 8.09 7.01 9.21 10.04 11.34 (0.41) (0.53) (0.27) (0.65) (0.46) (0.48) (0.46) (0.59) (0.41) (0.68) (0.74) (0.60) 28.37 (2.11) 20.79 (2.49) 29.10 (2.02) 7.34 (0.77) 29.08 (1.95) 34.98 (2.47) 175.80 145.45 71.20 26.77 136.05 98.80 Mongolians 5.66 7.79 4.92 8.10 7.75 10.79 7.63 8.16 6.67 9.04 9.80 11.46 32.66 21.11 28.22 7.49 29.33 35.45 176.80 135.65 70.38 25.42 134.10 102.60 (0.41) (0.56) (0.40) (0.72) (0.57) (0.49) (0.90) (0.61) (1.02) (1.04) (0.96) (0.65) (2.59) (1.19) (2.07) (0.61) (1.98) (2.23) (8.44) (5.40) (4.42) (1.68) (6.87) (5.31) Amerindians F= F= 2.58 2.70 4.55 3.65 3.60 2.58 3.28 1.31* 1.28* 4.96 3.88 1.81* 2.75 2.67 10.81 4.37 2.28* 2.86 8.69 3.41 F = 5.40 3.80 11.55 3.63 12.26 9.43 4.26 F = 7.49 F 6 6 in mml and uniuariate IF) and multiuatiate IWilks' A) variance ratios for one hundred human skulls.' Whites 182.20 141.65 66.34 23.48 128.15 100.65 ~~ Nontraditional craniometrics basioccipital length hasioccipital breadth posterior nares breadth carotid diameter foramen magnum breadth lacerum-jugular diameter Traditional craniometrics length breadth face height nasal breadth bizygomatic breadth basicranial distance ~__. TABLE 1. Basic descriptive statistics 0.530 0.469 0.365 0.130 Wilks' lambda U > z 126 D. FALK AND R. CORRUCCINI The univariate data for cervical dental measurements sort somewhat better than the maximum dental measurements. All of the F-ratios for cervical dentition were significantly greater than 1.0 while three of the six F-ratios for the maximum dental data were not significant at the 0.05 level. For both maximum and cervical dental measurements, the third premolar was the best separator of human groups (i.e., yielded the highest F-ratios). Multivariate analysis confirms the univariate results for cranial data. Traditional cranial measurements easily provide the best separation of human populations (Wilks’lambda = 0.130),followed by nontraditional cranial measurements (0.365). However, contrary to the univariate analysis, multivariate analysis of maximum dental measurements is slightly better at sorting human populations (Wilks’ lambda = 0.469) than cervical dental measurements (0.530). tage that maximum dental data have over cervical dental data when the data are subjected to multivariate analysis. Probably the higher multivariate partition of variance for maximum dimensions relates to a lower level of intercorrelation than is characteristic for cervical dimensions. Perhaps the field effect differs quantitatively, when studied by correlation, between maximum (enamel) and cervical (cemento-enamel junction) dimensions. We feel these results have implications for local microdifferentiation and microevolution, and substantiation should be sought in comparisons of closely related archeological populations. ACKNOWLEDGMENTS DISCUSSION AND CONCLUSIONS We thank Douglas Ubelaker of the United States Museum of Natural History (Smithsonian Institution) for hospitality and access to human osteological materials, John Karosas for key punching data, and Southern Illinois University for support for summer research (DF). Cranial measurements sort individuals in human populations better than do an equal number of dental measurements. We can think of several possible reasons for this: (1)Tooth information may be more redundant because of correlation. (2)Tooth size may be less heritable than generally thought (Townsend and Brown, 1978% b; Garn et al., 1979). (3) Perhaps teeth are less indicative of the major processes differentiating human races. (4)Teeth, because of relatively ill-defined landmarks, may be subject to greater measurement error. Within the cranial data, traditional measurements that span several bones proved to be better than nontraditional basicranial measurements that were each contained within one bone. It appears that the overall external shape of the skull may be more important in evolutionary terms than its constituent parts (including teeth). As emphasized by Howells (1969), many of the traditional cranial measurements have high value for sorting major human populations. Finally, our results suggest that cervical dental measurements are about as good for sorting human groups as the traditional maximum dental measurements. Thus, by collecting cervical dental data, odontologists and paleontologists can increase their sample sizes since there is no need to discard worn teeth as is the case when collecting maximum dental measurements (Black, 1978). Such increase in sample size might even offset the slight advan- Anderson, TW (1958)An Introduction to Multivariate Statistical Analysis. New York: Wiley. Black, TK (19781 A Method for Determining the Sex of Immature Human Skeletal Remains from Odontometric Dimensions. Ph.D. thesis, University of Michigan. Black, TK (1979) Dental wear and dental rnetrics. Am. J. Phys. Anthrop., 50:419-420 (abstract). Broca, P (1874) Memoires Danthropologie. Vol. 2 Paris: C Reinwald. Butler, PM (1939)Studies of the mammalian dentition: differentiation of the post-canine dentition. Roc . 2001. Soc. Lond.. 109B1-36. Corruccini, RS (1974) An examination of the meaning of cranial discrete traits for human skeletal biological studies. Am. J.Phys. Anthrop.. 40:425-446. Corruccini, RS (1977) Crown component variation in hominoid lower third molars. &it. Morph. Anthrop., a 1 4 - 2 5 . Dahlberg, AA (1945)The changing dentition of man. J. Am. Dent. Assoc.. 32:676-690. De Quatrefages. A, and Hamy. E 11882) Les Cranes des Races Humaines. Paris: J B Bailliere. Garn, SM, Osborne, R, and McCahe, K 11979) The effect of prenatal factors on crown dimensions. Am. J. Phys. A n throp., 51:665-678. Hanihara, K (1977) Distances between Australian aborigines and certain other populations based on dental measurernenls. J. Hum. Evol. 6403-418. Harris, EF, and Nweeia. M (1980)Tooth size of Ticuna Indians, Columbia, with phenetic comparisons to other Amerindians. Am. J. Phys. Anthrop.. 5381-91. Henderson, AM 11975)Dental field theory: an application to primate dental evolution. Ph.D. dissertation, University of Colorado, Boulder. Howells, WW (1969)Criteria for selection of osteometric dimensions. Am. J. Phys. Anthrop., 30:451-458. Hrdlia a . A (1907) Skeletal remains suggesting or attributed to earty man in North America. Bureau of American Ethnology. Bulletin 33, Washington, D.C.: Government Publishing Office. LITERATURE CITED CRANIAL AND DENTAL MEASUREMENTS Sofaer, JA, Bailit, HL and MacLean, CJ (1971) A developmental basis for differential tooth reduction in hominoid evolution. Evolution, 25509-517. Townsend, GC, and Brown, T (1978a)Heritability of permanent tooth size. Am. J. Phys. Anthrop., 49497-504. 127 Townsend. GC, and Brown, T (1978b) Inheritance of tooth size in Australian aboriginals. Am. J. Phys. Anthrop., 48305-314. Wolpoff, M H (1971) Interstitial wear. Am. J. Phys. Anthrop., 34205-228.