AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 62127-133 (1985) Diachronic Variation in Cranial Thickness of Near Eastern Populations PATRICIA SMITH, YOHANAN WAX, ADRIAN BECKER, AND SHMUL EINY Departments of Orthodontics (A.B.1 and Anatomy (PS., S. E.), Hadassah School ofDentaZ Medicine, Department of Statistics iY. W.),Hebrew University, Jerusalem, Israel KEY WORDS Near East, Cephalometrics, Cranial thickness, Diachronic change ABSTRACT Cephalometric radiographs were taken of 111 skulls of skeletal remains of populations living in Israel and Jordan during the last 12,000 years. From these radiographs, skull length and height, and cranial thickness were measured. For each sex and period, high correlations were found between cranial thickness at vertex, bregma, and lambda. Cranial thickness at nasion was correlated with sinus width but not sinus height. All measurements were correlated with skull length but not skull breadth. Using multivariate analysis, no significant differences in cranial thickness were found between the sexes. Significant diachronic trends were found in lambda and sinus width, and they were independent of variation in skull length. The Near East, with its long history of agriculture and pastoralism (Bar-Yosef, 19841, is a good region in which to examine the long-term effects of such adaptations on the human physique. Studies carried out on skeletal and dental morphology of Holocene populations have shown little evidence of unidirectional change in these parameters, but rather a series of fluctuations that have been interpreted as reflecting secular change and population replacement (HrdliEka, 1938; Arensberg, 1973; Ferembach, 1973; Smith et al., 1984). However, in a radiographic study, Smith et al. (1984) did find evidence of diachronic change in humeral cortical thickness over the last 12,000 years. We have accordingly begun investigating the total extent of variation in bone mass. In this report we present the findings of a radiographic study of cranial thickness in skulls of specimens ranging from the terminal Pleistocene (10,500 B.C.E.’) to recent (1,800 C.E.). In addition to chronological factors, a number of other variables have been associated with cranial thickness. Studies have been carried out both from direct measurement of skulls and cephalometric radiographs, the latter taken both of the living and of dry skulls. Such studies have shown significant 0 1985 ALAN R. LISS, INC. interpopulation variation not only in overall thickness, but also in relative thickness of frontal, parietal, and occipital bones. Adeloye et al. (1975) postulated that cranial growth patterns, as well as liability t o haemoglobinopathies affecting diploic width-especially of the parietals, might contribute to the differences observed. Brown et al. (1979) suggested that cranial thickness was associated with general skeletal robusticity, a hypothesis borne out by comparisons made by them, both of Australian crania of different periods and of Australian Aborigine and Caucasian populations. It has also been postulated that cranial thickness may be inversely related to cranial capacity (Todd, 1924)or to variation in piezoelectric currents associated with differences in the direction and intensity of magnetic fields operating on the earth’s surface a t different periods (Ivanhoe, 1979). In addition t o investigating population differences, a number of studies have been carried out on age-related changes in cranial thickness of the adult skull (Adeloye et al., 1975; Baer, 1956; Baer and Harris, 1969; Received July 15, 1983; revised March 26, 1984; accepted January 14,1985. ’B.C.E., Before Christian Era; C.E., Christian Era. 128 P. SMITH, Y.WAX, A. BECKER, AND S. EINY Hansman, 1966; Hellman, 1927; Israel, 1968, 1973; Tallgren, 1974). However, in two of the few longitudinal studies that have been carried out (Israel, 1968, 1973; Tallgren, 1974), conflicting conclusions were reached. Israel (1968) found a n overall increase of 5-6% in most dimensions measured, while Tallgren (1974) found no such changes. Since the available archaeological and anthropological evidence suggests several waves of population replacement in the Near East during the Holocene, any discussion of diachronic change in this region must include some measure of population differences as distinct from chronological sequence. In the present study we have attempted this through the incorporation of length and height measurements of the skull into the analysis. We have also grouped the specimens by biological age, as well as chronological sequence and sex, in order to test for the effects of these variables. Fig. 1. Tracing of cephalometric roentgenograph with lines constructed for measurement. MATERIALS AND METHODS Cephalometric roentgenographs were taken of 111skulls of skeletal remains excavated from a number of sites in Israel and Jordan. They were identified by period as follows: recent Bedouin, 1800 c.E.; Early Arab, 800-1500 c.E.; Late Hellenistic to Byzantine, 100 B.C.E.-400 c.E.; Middle Bronze Age, 2,200-1,550 B.c.E.; and Natufian, 10508500 B.C.E. For each individual, age and sex determinations were made from the entire skeleton, where present (after Krogman, 1962). Lateral cephalographic roentgenographs were then taken of the skulls, using a specially designed cephalostat, such that the cranium was inverted, as described by Brown (1973). To reduce differences in magnification of midline structure, the focus to midsagittal distance was fixed for each specimen a t 180 cm, and the midsagittal to cassette distance a t 15 cm. In addition a scale was suspended parallel to and opposite the midsagittal plane of each specimen and used to determine the magnification of midline structures on the roentgenograph. The roentgenographs so obtained were placed on a viewing box and covered with acetate paper. The skull was then traced onto the paper. In addition to anatomical landmarks, a number of constructed lines and angles were added to the tracing (Fig. 1). At bregma a line was drawn from the outer table of the skull, a t right angles to the Frank- fort plane. This line was used to measure skull height and cranial thickness a t vertex. Finally, maximum anterior-posterior length of the skull was measured along a line drawn parallel to the Frankfort plane. For the frontal sinus, maximum height and depth measurements were taken at right angles to one another, such that they incorporated the maximum superior-inferior and anterior-posterior dimensions of the sinuses (Fig. 2). All data were transferred to the computer. The measurements were corrected for magnification, so that analysis was carried out on corrected values. All measurements were taken by one observer, using Vernier calipers accurate to 0.1 mm, and the magnification was recorded from the scale, as projected on the film. Scattergrams and Pearson correlation coefficients were used to assess the relationship between the various cranial thickness parameters and between them and the corresponding length and height of the skull. The relationships were examined separately for each sex and age subpopulation in each period. Period differences in the seven cranial thickness measurements controlled for the effects of the confounding factors; sex, age, and skull length and height were estimated through the use of Multivariate Analysis of Covariance (MANCOVA) models (Hull and Nie, 1981).These models included period and sex factors, their interaction, and the covar- DIACHRONIC CHANGE IN CRANIAL THICKNESS 919 3 i N A B 129 I I I I I I I I C Fig. 2. Tracing of sinus outline to show variations and planes of measurement. iants-age (as a dichotomous variable), length-height, and sex-length. The goodness of fit of the model was examined by comparing the observed means of the cranial thickness measurements in period and sex subclass to the predicted means under the model. The significance of cranial thickness changes through time was assessed using the Roy union intersection test. Univariate analysis of the covariance model was used to test the significance of the changes through time in each of the cranial thickness characteristics. All analyses were performed using the SPSS-MANCOVA program, after Hull and Nie (1981). mmmmmmmnr oooooooc d M e i i i 3 - RESULTS The Natufian skulls were fragmentary, limiting the number of valid measurements that could be carried out on them. They were therefore omitted from the statistical analysis. Mean values for males were higher than those of females (Table 1)and increased with age (Table 2). Cranial thickness parameters showed positive, and on the average higher, correlation among males than among females (Table 3). The highest correlations found were between vertex and bregma (0.8 males, 0.68 females), and both were positively correlated with lambda (0.57 males, 0.32 and 0.36 females). A similar pattern applied to male-female values for each period studied, but Box's M-test for homogeneity of dispersion matrices revealed no significant differences (p = 0.15) between the dispersion matrices in period-sex subclasses. The correlation between skull length and cranial thickness was also fairly high but not that between skull height and cranial thickness. o o r n o o m o o mmNmmNmm 130 P. SMITH, Y.WAX, A. BECKER, AND S. EINY TABLE 2. Mean values and standard deuiation for skulls of differentperiods (sexes combined) S.D No X S.D No. X S.D No Middle Bronze X 1.6 1.5 1.0 4.7 2.4 3.4 8.7 5.6 24 24 23 23 23 24 24 24 7.6 5.8 5.9 20.4 7.3 12.1 176.1 115.0 1.2 1.5 1.2 5.6 2.0 2.9 7.7 5.1 15 15 15 15 15 15 15 15 6.5 5.6 5.5 23.3 9.8 14.3 174.1 113.4 1.7 2.1 2.1 5.4 2.6 2.7 9.8 5.7 14 14 14 12 12 13 14 11 6.8 5.8 6.1 22.2 9.2 13.1 181.8 113.0 1.6 1.2 1.1 5.6 2.5 2.7 7.6 3.7 2.5 2.7 1.0 6.5 1.2 1.8 3.7 4.7 12 12 12 12 12 12 12 12 8.7 6.3 6.5 23.5 9.7 14.8 179.7 116.7 0.9 0.8 1.3 6.3 2.5 4.8 7.0 6.4 9 9 9 9 9 9 9 9 7.2 5.1 5.0 25.4 10.3 14.7 176.6 113.1 1.8 1.2 1.3 6.0 2.5 4.0 8.6 4.7 2 2 7.3 8.2 7.9 27.2 10.0 11.7 183.4 135.5 1.7 3.5 3.7 5.3 1.2 1.7 11.3 53 Bedouin Parameter No X Age group 15-49 Lambda 19 7.9 5.6 Vertex 19 19 Bregma 6.0 18 Sinus height 20.5 Sinus width 18 7.9 19 Nasion 11.4 19 Length 178.6 19 Height 114.9 Age group 50 + Lambda 3 8.3 3 Vertex 6.0 Bregma 3 5.6 Sinus height 3 16.1 3 Sinus width 6.3 3 Nasion 12.7 3 Length 183.6 Height 3 116.3 - HellenisticByzantine Early Arab ~~ ~~ 2 2 2 2 2 2 S.D TABLE 3. Pearson correlation coefficients between skull characteristics males (aboue diagonal) and females (below diagonal) Lambda Vertex Bregma Sinus height Sinus width Nasion Length Height Lambda Vertex Bregma - 0.57 0.57 0.80 0.32 0.36 0.19 0.02 0.04 0.16 0.02 I 0.68 0.06 0.23 -0.22 0.42 0.06 0.07 0.10 -0.04 0.49 0.05 This was particularly true for males. In view of these findings, length, and its interaction with sex, were introduced as covariants in subsequent analyses. In view of the slight differences found between younger and older age groups, age was also incorporated as a covariant in the multivariate model. Table 4 shows the results of the covariance model, with cranial thickness as the dependent variable, period and sex as grouping factors, and covariates as previously defined. As the significance levels show, only period showed a significant effect (p = 0.007), and there was no significant interaction between sex and period. Values obtained for individual parameters are shown in Table 5. Significant differences over time were found only for lambda and sinus width. The observed, expected, and adjusted means calculated for these two variables are shown in Figure 3, with the multiple comparison taken into ac- Sinus Ht Sinus Wdth 0.18 0.14 0.17 0.32 0.32 0.27 0.67 0.50 0.27 -0.07 -0.03 0.25 0.09 0.14 Nasion Length Height 0.39 0.28 0.24 0.46 0.71 - 0.41 0.40 0.36 0.17 0.27 0.29 -0.05 -0.14 0.28 0.23 0.26 -0.01 0.02 0.08 0.34 0.43 - - TABLE 4. Multivariate significance tests for overall differences in cranial thickness parameters among the factor levels Factor Period Sex Period bv sex interaction p-value for Roy test 0.007 0.97 0.19 count. The small differences between observed and expected means demonstrates the validity of the model used. The adjusted means for lambda increased while those for sinus width decreased. Major significant differences were present between the two more recent groups (Bedouin and Arab) and the two earlier groups (Hellenistic and Bronze Age). Although male and female adjusted values for lambda width were widely separated, no significant differences were present (p = 0.5). 131 DIACHRONIC CHANGE IN CRANIAL THICKNESS isticByzantine skulls, and one Middle Bronze Age skull fell into this category. They included females and both young and old individuals in whom cranial thickness at vertex ranged between 7.4 and 9.4mm. In no instance was any pathological lesion such as Paget's disease or haemoglobinopathies identified. The Natufian skulls were larger than all others measured, and thicker at vertex and bregma (Table 1).They resembled the Middle Bronze crania most closely. TABLE 5. Llniuariate F-test' for differences over time F D-value 4.59 0.86 1.27 1.59 4.58 2.02 0.005 0.47 0.29 0.20 0.005 0.12 Parameter Lambda Vertex Bregma Sinus height Sinus width Nasion 'With (3.81)degrees of freedom. Cranial thickness at vertex and bregma, while not showing statistically significant differences between periods, was thicker in the earlier, Middle Bronze Age sample than in other samples studied. This was so despite the fact that a relatively larger percentage of Early Arab skulls had exceptionally thick cranial bones. Ten skulls were exceptionally thick, with cranial thickness at vertex more than one standard deviation greater than the mean value of the total sample. Seven Early Arab skulls, two Bedouin skulls, two Hellen- DISCUSSION The high correlation found between the variables studied here demonstrates the importance of a multivariate approach to the analysis of diachronic trends. However, such studies are frequently subject to limitations imposed by sample sizes. In the present study this restricted the number of age categories used. In order to best control for age effects, tests were carried out twice, once using ages SINUS Y I D T H 10- 9B- I 7- 8 6- w A .d 54- w d .* 3- 2- 8 1- 8 0 I I I LWBDA ld9B- x ?65- o A . 4 x m m A * b o 4- L 32- 15 O A I X Ii "I I 0 D x L 0 I I 57 69 Fig. 3. Observed, expected, and adjusted values for mean values of lambda and frontal sinus width. Horizontal axis = measurements in mmx 10; vertical axis = group, where 1,3,5, ,7 represent males of increasing age from Bedouin to Middle Bronze and 2, 4, 6, 8 represent 81 93 Lm females ranked from Bedouin to Middle Bronze; open triangles = adjusted values; open squares = observed values; crosses = expected values; closed squares = superimposition of observed and expected values. 132 P. SMITH, Y. WAX, A. BECKER, AND S. EINY 15-39 and 40+ and once using ages 15-49 and 50+. The results obtained were similar for both age groupings and not significant. Our findings regarding the lack of statistically significant sex differences in cranial thickness a t any location other than nasion, on uncorrected data agree with those of Brown et al. (1974) on Australian Aborigines and Adeloye et al. (1975) on North American whites and blacks. In our study, as in the study by Brown et al. (1979) a tendency was found for female values to be smaller than those of males, although Adeloye et al. reported a tendency for increased frontal bone thickness in older females. This is a n interesting finding in view of the dimorphism shown by all the groups studied here, in length and height of the skull. It is also interesting to note that sex differences in the development of muscle markings in the nuchal crest region (Krogman, 1962) are not reflected in cranial thickness at lambda. The pronounced interaction shown between skull length and thickness of cranial bones further suggests that shorter skulls are relatively thicker. Age-related changes in cranial thickness were similarly insignificant in our samples. While this study was, by necessity, a crosssectional study, reference to other studies supports the hypothesis that thickness of cranial bones shows little correlation with age. Todd (1924) and Adeloye et al. (1975) both found age-related oscillations in cross-sectional samples studied by them. Todd’s study, like ours, was based on dry skulls, while Adeloye et al. measured cephalographs taken of living individuals. The longitudinal studies, by Israel et al. (1968, 1973) and Tallgren (1974), yielded conflicting results. This may be related to differences in the age groups selected by them for study. All of Israel’s subjects were younger than 56 years, at the time of the second phase of the longitudinal study, while Tallgren’s (1974) subjects COVered a wider (20-73 years) age range. As expected from reference to published studies on the skeletal morphology of Near Eastern populations (Arensberg, 1973; Ferembach, 1976; Smith et al., 19841, the Bronze Age to recent populations showed fluctuations in skull length and height ratios suggestive of population replacement rather than unidirectional trends. Diachronic changes were present a t lambda and in sinus width. Cranial thickness a t lambda was thickest in the two most recent groups and was independent of skull length, while sinus width was greatest in the two earliest groups. However, a marked difference in cranial thickness at vertex and bregma and in sinus size was also found between the Natufian and all later groups. A tendency was found for the Middle Bronze Age group to approach the Natufians most closely in these parameters. Comparison with populations studied elsewhere is limited because of differences in measurement techniques used. Israel (1968, 19731, Tallgren (19741, and Adeloye et al. (1975) used constructed landmarks rather than anatomical ones, as used here, and/or failed to correct for radiographic enlargement, which averages 6-8% for midline structures using conventional cephalometric techniques. Cranial thickness at lambda, vertex, and bregma in Near Eastern Middle Bronze and later periods lie well within the range of variation of modern European and North American Caucasians, and is thinner than that of North American blacks or Australian Aborigines. Nasion is relatively thick in our samples, and the frontal sinus is large. This appears to be a characteristic of recent Caucasians as compared with other population groups. ACKNOWLEDGMENTS This work was supported by a grant from the Israel Academy of Sciences. Part of this study is based on work carried out by S. Einy in partial fulfillment of the D.M.D. requirements of the Hebrew University. The authors would like to thank Y. Ainan for carrying out data analysis, J. Berkowitz for her painstaking work with the illustrations, and Hannah Siegal for coping with the typing and numerous tables. We are particularly grateful to an anonymous reviewer for his constructive and detailed advice during the review stage of this work. LITERATURE CITED Adeloye, A, Kattan, K, and Silverman, FN (1975)Thickness of the normal skull in American blacks and whites. Am. J. Phys. Anthropol. 43t23-30. Arensburg, B (1973)The people of Israel from Epipaleolithic times to the present. Ph.D thesis, University of Tel Aviv. Baer, M J (1956) Dimensional changes in the human head and face in the third decade of life. Am. J. Phys. Anthropol. 14.557-576. Baer, MJ, and Harris, JE (1969) A commentary on the growth of the human brain and skull. Am. J. Phys. Anthropol. 3Ot39-44. DIACHRONIC CHANGE IN CRANIAL THICKNESS Bar Yosef, 0 (1984) The Mediterranean Levantine EpiPaleolithic as the background to the Neolithic Revolution. In P Sorenson and P Mortensen (ed):The Origins of Agriculture and Technology: West or East Asia? University of Aarhus Press (in press). Brown, T, Pinkerton, SK, and Lambert, W (1979)Thickness of cranial vault in Australian Aboriginals. Arch. Phys. Anthropol. Oceania. 14~54-71. Ferembach, D (1973) L’evolution humaine au Proche Orient. Paleorient 1213-21. Hansman, CF (1966) Growth of interorbital thickness and skull thickness as observed in roentgenographic measurement. Radiology 86t87-96. Hellman, M (1927)Changes in human face brought about by development. 1st Int. Orth. Cong. New York, pp. 80-121. HrdliEka, A (1938) Skeletal remains in Megiddo tombs. PLO Guy and RM Engberg (eds):O.I.P. 32~192-208. Hull, CH, and Nie, NH (1981) SPSS Update 7-9. New York: McGraw-Hill. Israel, H (1968) Continuing growth in human cranial 133 skeleton. Arch. Oral. Biol. 13:133-137. Israel, H (1973)The dichotomous pattern of craniofacial expansion during aging. Am. J. Phys. Anthropol. 47:47-51. Ivanhoe, F (1979)Direct correlation of human skull vault thickness with geomagnetic intensity in some northern hemisphere populations. J. Hum. Evol. 8:433-444. Krogman, WM (1962) The human skeleton in forensic medicine. Springfield, IL: Charles C. Thomas. Smith, P, Bar Yosef, 0, and Sillen, A (1984)Archeological and skeletal evidence for dietary change during the late Pleistocene/Early Holocene in the Levant. In M Cochen and G Armelagos (eds): Paleopathology and Agricultural Origins. Academic Press: New York pp. 101-136. Tallgren, A (1974)Neurocranial morphology and aging: A longitudinal Roetgen, cephalometric study of adult Finnish women. Am. J. Phys. Anthropol. 41~285-293. Todd, TW (1924) Thickness of the male white cranium. Anat. Rec. 27:245-256.