Dental fluctuating asymmetry in the Gullah Tests of hypotheses regarding developmental stability in deciduous vs. permanent and male vs. female teethкод для вставкиСкачать
AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 129:427–434 (2006) Dental Fluctuating Asymmetry in the Gullah: Tests of Hypotheses Regarding Developmental Stability in Deciduous vs. Permanent and Male vs. Female Teeth Debbie Guatelli-Steinberg,1* Paul W. Sciulli,1 and Heather H.J. Edgar2 1 Department of Anthropology and Department of Evolution, Ecology, and Organismal Biology, Ohio State University, Columbus, Ohio 43210-1364 2 Maxwell Museum of Anthropology and Department of Anthropology, University of New Mexico, Albuquerque, New Mexico 87131 KEY WORDS Gullah; dentition; teeth; asymmetry; developmental stability ABSTRACT In this investigation, deciduous teeth (canines, c; ﬁrst molars, m1; second molars, m2) and their permanent successors (canines, C; ﬁrst premolars, P1; second premolars, P2) were used to test two related hypotheses about ﬂuctuating asymmetry (FA). First, based on the biology of the developing dentition, it was predicted that deciduous teeth would be more developmentally stable and thus exhibit less dimensional FA than their permanent successors. Second, based on sex differences in tooth development, it was predicted that female canines would have greater developmental stability (less FA) than male canines. Bucco-lingual measurements were made on dental casts from a single Gullah population. Using a repeated-measures study design (n ¼ 3 repeated measures), we tested these hypotheses on sample sizes ranging from 63–82 antimeric pairs. Neither hypothesis was supported by our data. In most cases, Gullah decidu- ous teeth did not exhibit statistically signiﬁcantly less FA than their permanent successors; indeed, statistically signiﬁcant differences were found for only 3 of 12 deciduous vs. permanent contrasts, and in two cases, the deciduous tooth had greater FA. Female mandibular canines exhibited statistically signiﬁcantly greater FA than those of males, while there was no statistically signiﬁcant sex difference in FA for the maxillary canine. FA in these Gullah samples is high when compared to Archaic and late prehistoric Ohio Valley Native Americans, consistent with historical and archaeological evidence that environmental stress was relatively higher in the Gullah population. We suggest that when environmental stress in a population is high, the impact of differences in tooth formation time spans and developmental buffering upon FA may be minor relative to the effect of developmental noise. Am J Phys Anthropol 129:427–434, 2006. V 2005 Wiley-Liss, Inc. In bilaterally symmetric organisms, ﬂuctuating asymmetry (FA) can be deﬁned as asymmetry ‘‘that results from the inability of organisms to develop in precisely determined paths’’ on both sides of the body (Van Valen, 1962). This inability reﬂects a compromise between two sets of processes: developmental noise, ‘‘a suite of processes that tend to disrupt precise development’’ (Palmer, 1994, p. 337), and developmental stability, ‘‘a suite of processes that tend to resist or buffer the disruption of precise development’’ (Palmer, 1994, p. 337). Factors that may increase developmental noise include, but are not limited to, extreme temperature, audiogenic stress, malnutrition and disease, congenital defects, chromosomal defects, maternal alcohol consumption, maternal obesity, maternal smoking, congenital defects, and possibly also inbreeding (e.g., Bailit et al., 1970; Kieser, 1992; Kieser et al., 1997; Siegel and Doyle, 1975a,b; Siegel et al., 1977; Sciulli et al., 1979; Narayanan et al., 1999; Niswander and Chung, 1965; Townsend, 1983; Townsend and Garcia-Godoy, 1984). In contrast to what is known about causes of developmental noise, current knowledge regarding the physiological processes underlying developmental stability is much more limited (Lens et al., 2002), although Quetisch et al. (2002) identiﬁed a cellular protein (heat-shock protein 90) that stabilizes other cellular proteins under conditions of external stress. In a sample of individuals, ﬂuctuating asymmetry is manifested as a pattern of ‘‘bilateral variation . . . where the mean of R L [right minus left] is zero and variation is normally distributed about that mean’’ (Palmer, 1994, p. 338). It differs from directional asymmetry in which there is normally ‘‘greater development of a character on one side of the plane of symmetry . . . than the other’’ (Van Valen, 1962, p. 125). Van Valen (1962) cited the mammalian heart as an example of directional asymmetry. Fluctuating asymmetry also differs from antisymmetry in which there is ‘‘a negative correlation, usually implying a negative interaction, between the development of a character on the two sides’’ (Van Valen, 1962, p. 126). For antisymmetry, there is a consistent mean difference between sides, but the ‘‘side that is larger varies at random among individuals’’ (Palmer 1994, p. 337). Fluctuating asymmetry is of particular interest to biologists and anthropologists as an indicator of environmental stress in populations (reviewed in Lens et al., 2002; Kieser, 1990; Fraser, 1994). Lens et al. (2002, p. 27) found that FA does not ‘‘consistently index stress or ﬁtness’’ in animal populations. Fraser (1994, p. 319) similarly C 2005 V WILEY-LISS, INC. C *Correspondence to: Dr. Debbie Guatelli-Steinberg, Department of Anthropology, Ohio State University, 124 West 17th Ave., 244 Lord Hall, Columbus, OH 43210-1364. E-mail: email@example.com Received 8 October 2004; accepted 16 November 2005.. DOI 10.1002/ajpa.20237 Published online 1 December 2005 in Wiley InterScience (www.interscience.wiley.com). 428 D. GUATELLI-STEINBERG ET AL. reviewed studies of FA in human dental dimensions, stating that these studies provide no ‘‘convincing evidence’’ that stressful environments increase FA. Lens et al. (2002) found that the lack of consistency across studies is in part the result of conceptual issues (e.g., our limited understanding of the mechanisms underlying developmental stability) and in part the result of statistical issues involved in comparing FA levels between or among groups. FA differences between samples are assessed by testing for differences in variances where small differences (e.g., F ¼ 1.25 at ¼ 0.05) are not likely to be detected unless sample sizes are large. Smith et al. (1982) found that sample sizes of several hundred are required for this variance ratio, while Kieser (1990) argued that sample sizes of 75 or more are adequate. Sample sizes smaller than this may have led to type II errors in some studies attempting to test the relationship between FA and environmental stress. In this paper, we examine two additional factors complicating the use of dental FA as a stress indicator in humans: differences in permanent vs. deciduous and male vs. female tooth formation. Differing conditions speciﬁc to each (deciduous vs. permanent, and male vs. female) may cause them to differ in the degree to which they are buffered against growth disturbances. First, because deciduous tooth crowns take less time to form than permanent tooth crowns, both in their soft-tissue and mineralization stages (e.g., Mizoguchi, 1980, 1998), and may also be buffered against developmental disturbances by the intrauterine environment in which they grow (Cook and Buikstra, 1970; Sciulli, 1978), we predicted that deciduous teeth would exhibit less FA than their permanent successors. Second, female permanent canine teeth were expected to show less FA than those of males for various reasons: females may be better buffered against developmental disturbances (Stinson, 1985), paired X-chromosomes may confer on females greater dimensional control during odontogenesis (Garn et al., 1965, 1966), and the permanent canine teeth of females, because they mineralize in relatively less time (Moss, 1978; Moss and Moss-Salentijn, 1976), may have a smaller window of opportunity for developmental disturbances to occur (Townsend and Farmer, 1998; Townsend, 1981). The period during which developmental disturbances can have their greatest impact on tooth size begins with the formation of the enamel organ and ends during the bell stage, when mineralization begins (Harris, 2002). Thus, the duration of the soft-tissue stage of development is especially relevant. Previous research did not ﬁnd consistent dental FA difference between deciduous and permanent dentitions or between males and females. Research on FA in deciduous vs. permanent teeth suggests that the former exhibits less FA than the latter for dental morphological traits (Saunders and Mayhall, 1982), but Hershkovitz et al. (1993) found only two cases out of 12 comparisons in which permanent and deciduous teeth differed signiﬁcantly in metric FA, and in both cases, the deciduous teeth had the greater mean rank (Kruskal Wallis one-way ANOVA). Garn et al. (1966) and Townsend and Brown (1980) found males to have greater FA in their permanent dentition than females; Harris and Nweeia (1980) found greater FA in females; and Kieser et al. (1986) did not ﬁnd a statistically signiﬁcant sex difference. The inconsistent ﬁndings of the above studies may represent real population differences but are also likely to reﬂect different methods of analysis, and the fact that not all studies achieved adequate sample sizes, tested for anti- symmetry, or partitioned out measurement error and directional asymmetry from the between-sides variance. Here, we reexamine the question of permanent vs. deciduous and male vs. female differences in dental FA by taking these complicating factors into account in our statistical analysis. We thus test for antisymmetry and adjust our FA estimates for directional asymmetry, size or shape variation, and measurement error. MATERIALS AND METHODS Sample The data derive from dental casts of a single population of Gullah, African Americans from the Outer Banks of South Carolina, living on St. James Island during the 1950s (Menegaz-Bock, 1968). The dental casts (N ¼ 469) were made as part of a larger study of Gullah biology and ancestry (Menegaz-Bock, 1968), and are now housed in the Renée M. Menegaz-Bock Dental Anthropology Collection of Ohio State University. Living anthropometrics as well as genealogies are available for the collection. Individuals included in this study were selected to minimize their interrelatedness: only one member of each group of related individuals was included. Antimeric pairs of the following teeth were measured: maxillary and mandibular deciduous canines (c), ﬁrst molars (m1), second molars (m2), and their permanent successors (canines, C; ﬁrst premolars, P1; second premolars, P2). These particular teeth were chosen because there were few dentitions with deciduous anterior teeth. Tables 1–3 give sample sizes. For comparative purposes, we also included data from a previous study of dental asymmetry in Native American Late Archaic hunter-gatherers and late prehistoric maize agriculturalists of the Ohio Valley (Sciulli, 2002). Details and sample sizes for these groups are mentioned in Sciulli (2002), so they are not reproduced here. The data from these groups provide a context for evaluating the magnitude of dental asymmetry in the Gullah. Measurement Bucco-lingual (BL) diameters were measured using Mitutoyo digital calipers (instrument accuracy ¼ 0.01 mm). Buccolingual diameters are the ‘‘distances between points of maximum curvature of the buccal and lingual surfaces of a tooth’’ (Sciulli, 2002) and are measured perpendicularly to the mesio-distal axis (Kieser, 1990). Teeth were excluded from the study if excessive wear or carious lesions affected the maximum BL diameter. D.G.-S. measured each tooth three times on separate occasions, with intervals of at least 1 week between them. P.W.S. reviewed these measurements, noting recording errors and perceived measurement errors. D.G.-S. then remeasured the teeth in question without knowing the previous measures or the exact nature of the error. These corrections were then used in the analysis. Statistical analyses Descriptive statistics were generated for each tooth type: mean size, standard deviation, coefﬁcient of variation, and measurement error. Measurement error is (MSM)1/2, where MSM is the error mean square from a sides individual ANOVA (Palmer, 1994). To analyze dental asymmetry, we used a two-way, mixed-model analysis of variance (ANOVA) with repeated measures, where sides are ﬁxed and individuals are ran- 429 GULLAH FLUCTUATING DENTAL ASYMMETRY TABLE 1. Sample sizes and descriptive statistics: Gullah deciduous teeth Left Right Tooth N Mean S CV Mean S CV Measurement error (mm)1 Max C Max m1 Max m2 Mand c Mand m1 Mand m2 68 67 63 67 65 63 5.78 8.21 9.56 5.35 6.72 8.67 0.44 0.44 0.44 0.39 0.50 0.49 7.57 5.33 4.63 7.24 7.43 5.63 5.72 8.29 9.56 5.29 6.65 8.75 0.49 0.47 0.49 0.44 0.52 0.51 8.59 5.66 5.10 8.37 7.85 5.79 0.13 0.16 0.15 0.12 0.19 0.15 1 From side individual ANOVA; (mean square measurement)1/2, where mean square measurement ¼ error mean square. Max ¼ maxillary, Mand ¼ mandibular. TABLE 2. Sample sizes and descriptive statistics: Gullah female permanent teeth Left Right Tooth N Mean S CV Mean S CV Measurement error (mm)1 Max C Max P1 Max P2 Mand C Mand P1 Mand P2 82 79 72 73 81 73 8.09 9.75 9.75 7.24 8.10 8.63 0.57 0.57 0.59 0.59 0.56 0.61 7.09 5.81 6.01 8.09 6.97 7.02 8.12 9.74 9.74 7.23 7.98 8.53 0.55 0.58 0.60 0.57 0.55 0.64 6.83 5.97 6.17 7.84 6.83 7.54 0.10 0.09 0.12 0.12 0.13 0.13 1 From side individual ANOVA; (mean square measurement)1/2, where mean square measurement ¼ error mean square. Max ¼ maxillary, Mand ¼ mandibular. TABLE 3. Sample sizes and descriptive statistics: Gullah male permanent teeth Left Right Tooth N Mean S CV Mean S CV Measurement error (mm)1 Max C Max P1 Max P2 Mand C Mand P1 Mand P2 74 82 71 70 81 69 8.60 9.93 9.90 7.75 8.42 8.82 0.61 0.65 0.60 0.64 0.63 0.60 7.14 6.50 6.03 8.21 7.53 6.82 8.54 9.96 9.93 7.71 8.29 8.71 0.63 0.62 0.58 0.60 0.62 0.64 7.42 6.17 5.85 7.80 7.51 7.31 0.10 0.14 0.13 0.13 0.13 0.17 1 From side individual ANOVA; (mean square measurement)1/2, where mean square measurement ¼ error mean square. Max ¼ maxillary, Mand ¼ mandibular. dom (Palmer and Strobeck, 1986). Table 4 contains the ANOVA model with the expected mean squares, in which S is the number of sides, J is the number of individuals, and M is the number of measures per side. Three kinds of asymmetry were considered: antisymmetry, directional asymmetry, and ﬂuctuating asymmetry (Van Valen, 1962). In the ANOVA with one individual per genotype, the contribution of antisymmetry to nondirectional asymmetry cannot be estimated. In extreme cases, antisymmetry results in a bimodal distribution of the signed difference between sides, while in more subtle cases, the distribution will tend toward platykurtosis. We assessed antisymmetry by testing for the normality of the (R L) distributions, including tests for kurtosis and skewness (D’Agostino, 1986; Zar, 1999). Fluctuating asymmetry was evaluated by the F statistic, using the remainder mean square (MSSJ) over the mean square due to measurement error (MSM). If signiﬁcant, the variance component due to nondirectional asymmetry is i2 ¼ (MSSJ MSM)/3, where 3 is the number of repeated measures per tooth. Nondirectional asymmetry contains contributions from antisymmetry and ﬂuctuating asymmetry. The presence of antisymmetry is evaluated by the normality test, and if absent, leaves only ﬂuctuating asymmetry in the nondirectional component. Size and shape variation were eliminated as sources of variation by dividing each side by (Ri þ Li)/2. Thus the sum of squares and mean squares for the individual source of variation is in all cases zero. Directional asymmetry was evaluated using the mean square for sides (MSS) over the remainder mean square (MSSJ). We also calculated the repeatability for each tooth in order to investigate how much accuracy would be gained by repeated measurements (Falconer and MacKay, 1997). The ANOVA model for repeatability is given in Table 5. RESULTS Descriptive statistics Descriptive statistics (mean, standard deviation (SD), coefﬁcient of variation (CV), and measurement error) are given in Tables 1–3. Because there were too few deciduous teeth of each sex to conduct and analysis by sex and because deciduous teeth are compared to permanent teeth in the asymmetry analysis, the descriptive data for deciduous teeth are presented for both sexes combined. There were approximately equal numbers of each sex in the sam- 430 D. GUATELLI-STEINBERG ET AL. TABLE 4. ANOVA model for evaluating asymmetry Source of variation Sides Individuals Remainder Measures 1 Degrees of freedom1 Mean square Expected mean square Interpretation (S 1) (J 1) (S 1)(J 1) SJ(M 1) MSS MSJ MSSJ MSM 2 m þ M(i2 þ (J/S 1)S2) 2 m þ M(i2 þ Sj2) 2 m þ Mj2) 2 m Directional asymmetry Size/shape variation Nondirectional asymmetry Measurement error S ¼ 2, J ¼ number of individuals, M ¼ 3 (measures per side). TABLE 5. ANOVA model for evaluating repeatability Source of variation Degrees of freedom1 Among individuals Within individuals J1 J (M 1) 1 Mean square MSB MSW W2 W2 þ MB2 Repeatability B2/(B2 þ W2) B2 ¼ (MSB MSW)/6. TABLE 6. Repeatability values for deciduous and permanent (females, males) teeth1 Tooth Deciduous (SE) Maxillary canine Maxillary P1 or m1 Maxillary P2 or m2 Mandibular canine Mandibular P1 or m1 Mandibular P1 or m2 0.87 0.84 0.85 0.86 0.83 0.86 1 Expected mean square (0.03) (0.03) (0.03) (0.03) (0.03) (0.03) Females (SE) 0.92 0.96 0.92 0.88 0.87 0.89 (0.02) (0.01) (0.02) (0.02) (0.02) (0.02) Males (SE) 0.92 0.92 0.91 0.88 0.89 0.88 (0.02) (0.02) (0.02) (0.02) (0.02) (0.02) SE, standard error. ple of deciduous teeth. In general, mandibular teeth, both permanent and deciduous, show greater variation (CVs) than their maxillary counterparts. Except for the mandibular m2, the relative variation of deciduous teeth is as great as their permanent successors. Measurement error ranges between 0.09–0.19 mm. Measurement error as a percentage of mean BL diameter ranges between 0.9% (female maxillary right P1) and 2.9% (mandibular left m1). Our tests for repeatability (Table 6) revealed that with our degree of measurement error, we could not have increased our accuracy by taking more than three measurements (Falconer and Mackay, 1997). Normality We performed 18 tests of normality for the (R L) distributions of the deciduous teeth (three for each tooth). Thirty-six such tests were performed for the permanent teeth (three for each tooth for each sex). For the deciduous teeth, there was skewness in measure 2 of the mandibular m1 (P ¼ 0.022) and measure 1 of the mandibular m2 (P ¼ 0.010). Skewness in these measures results from some individuals exhibiting extreme values. These outliers were checked and kept in the data set. None of the teeth showed consistent deviation for all three measures. Importantly, none of the deciduous distributions were platykurtic, indicating that antisymmetry was not present. For the permanent teeth, there was no skewness, although ﬁve measures were slightly leptokurtic: female maxillary P2, measure 1 (P ¼ 0.030) and measure 2 (P ¼ 0.0220); female mandibular P1, measure 1 (P ¼ 0.033); male maxillary P1, measure 3 (P ¼ 0.010); and male maxillary P2, measure 1 (P ¼ 0.040). Graphical inspection of these data revealed that outliers caused the leptokurtosis. These outliers were checked and retained in the data set. Again, in the permanent dentition, there was no platykurtosis in (R L) distributions, indicating that there was no antisymmetry. Directional and ﬂuctuating asymmetry T-tests of the means of (R L) distributions indicate that the means are signiﬁcantly different from zero for ﬁve deciduous teeth and four permanent teeth. Thus these teeth showed directional asymmetry. In the deciduous dentition, right was greater than left in the maxillary canine, mandibular canine, and mandibular m1, while left was greater than right in the maxillary m1 and mandibular m2. In the permanent dentition, right was greater than left for both male and female mandibular P1 and P2. However, in all cases, the mean (RL) was 2–6 times smaller than 0.798 (variance (RL))1/2, indicating that the predisposition toward one side is less than the average deviation about the mean (RL). (If developmental instability varies among individuals, the distribution of RL is no longer normal, and the expected value (mean) of the distribution differs from the expected standard deviation by (2/)1/2 ¼ 0.798.) Directional asymmetry is thus unlikely to confound interpretations about FA variation, as deviations about the mean directional asymmetry are largely due to developmental instability (Palmer and Strobeck, 2003). Results of ANOVA tests are given in Tables 7–9. The size-correction transformation eliminates the size-shape (individual) source of variation (MSJ ¼ 0). Thus, only three sources of variation remain: sides, ﬂuctuating asymmetry (this is equal to nondirectional asymmetry because antisymmetry is not present), and measurement error. As seen in Tables 7–9, there is signiﬁcant nondirectional as well as directional asymmetry (but see above) in both dentitions and in both males and females. Here, we list the teeth from those with most to least FA and group within parentheses those that do not differ signiﬁcantly from each other, as revealed by F-tests. For the deciduous dentition, the order from greatest to least FA (i2) is: (maxillary canine, mandibular canine, and mandibular m1) and (maxillary m1, maxillary m2, and mandibular m2). For the female permanent dentition, the order from greatest to last FA is: (mandibular C (mandibular P2, mandibular P1) (maxillary canine) maxillary P2) (maxillary P1). For the male permanent dentition, FA from greatest to least is: (mandibular P1, mandibular P2, maxillary canine (mandibular canine) maxillary P2, and maxillary P1). Figure 1 plots the variance component due to FA (i2 104) for each tooth type. Enclosed in each rectangular box 431 GULLAH FLUCTUATING DENTAL ASYMMETRY TABLE 7. Two-way mixed-model ANOVA: Gullah deciduous teeth with 3 repeated measures per side1 Tooth N xc 68 xml xm2 dc dml dm2 1 2 3 4 5 6 7 67 63 67 65 63 Source of variation2 DF MS3 F4 i2 DF5 Sides Remainder Measurement Sides Remainder Measurement Sides Remainder Measurement Sides Remainder Measurement Sides Remainder Measurement Sides Remainder Measurement 1 67 272 1 66 268 1 62 252 1 66 268 1 64 260 1 62 260 127.25 23.30 2.30 113.20 10.78 1.53 0.44 0.50 9.37 166.91 23.10 2.25 107.37 21.22 4.50 76.50 10.77 1.04 5.466 10.11 7.00 54 10.50 7.03 3.08 51 0.05 9.60 2.80 61 7.226 10.25 6.95 53 5.066 4.71 5.57 39 7 7.10 10.30 TABLE 9. Two-way mixed-model ANOVA: Gullah female permanent teeth with three repeated measures per side1 Tooth N XC 82 XP1 XP2 DC DP1 7 DP2 3.24 48 1 x, maxillary; d, mandibular. Individual source of variation ¼ 0 with size correction. 104. Boldface indicates P < 0.05. Sheffé (1959). R > L. L > R. 2 3 4 5 6 79 72 73 81 73 Source of variation2 DF MS3 F4 i2 DF5 Sides Remainder Measurement Sides Remainder Measurement Sides Remainder Measurement Sides Remainder Measurement Sides Remainder Measurement Sides Remainder Measurement 1 81 328 1 78 316 1 71 288 1 72 292 1 80 324 1 72 292 12.47 14.06 0.56 0.57 4.85 0.36 1.97 10.21 0.47 0.84 27.69 0.97 290.95 17.80 0.80 132.90 20.16 0.99 0.89 25.14 4.50 74 0.11 13.47 1.50 66 0.19 21.62 3.24 61 0.03 28.60 8.91 67 16.356 22.13 5.67 73 6.59 20.41 6.39 65 6 X, maxillary; D, to mandibular. Individual source of variation ¼ 0 with size correction. 104. Boldface indicates P < 0.05. Sheffé (1959). R > L. TABLE 8. Two-way mixed-model ANOVA: Gullah male permanent teeth with three repeated measures per side1 Tooth N XC 74 XP1 XP2 DC DP1 DP2 1 2 3 4 5 6 82 71 70 81 69 Source of variation2 DF MS3 F4 i2 DF5 Sides Remainder Measurement Sides Remainder Measurement Sides Remainder Measurement Sides Remainder Measurement Sides Remainder Measurement Sides Remainder Measurement 1 73 296 1 81 328 1 70 284 1 69 280 1 80 324 1 68 276 63.17 15.24 0.61 16.65 9.27 0.99 6.42 9.11 0.62 29.95 15.04 1.06 305.27 19.06 0.73 153.85 16.38 1.50 4.15 25.01 4.88 66 1.80 9.33 2.76 66 0.70 14.61 2.83 59 1.99 14.12 4.66 44 16.016 26.14 6.11 75 9.396 10.96 4.96 56 X, maxillary; D, mandibular. Individual source of variation ¼ 0 with size correction. 104. Boldface indicates P < 0.05. Sheffé (1959). R > L. are values of i2 that are not statistically signiﬁcantly different from each other as assessed by F-tests, given in Table 10. There are 12 comparisons of deciduous teeth and permanent teeth: each of the six deciduous teeth is compared to the six male and six female permanent counterparts. Of the 12 deciduous vs. permanent comparisons (Table 10), the permanent successor has a greater FA component than the deciduous tooth in seven cases, with one compar- Fig. 1. Bilateral variance component due to FA (i2 104) for each tooth type. Values enclosed in rectangles do not differ signiﬁcantly from each other, as assessed by F-tests shown in Table 10. GullahD, Gullah deciduous; GullahPF, Gullah permanent female; GullahPM, Gullah permanent male; LP, late prehistoric Ohio Native Americans; LA, Late Archaic Native Americans; X, permanent maxillary; D, permanent mandibular; x, deciduous maxillary; d, deciduous mandibular. ison reaching statistical signiﬁcance (mandibular m2female P2). Of the ﬁve cases in which the deciduous tooth has the greater FA component, two reach statistical significance (deciduous maxillary canine female maxillary canine and maxillary m1-female P1). Three comparisons in Table 10 resulted in F-values near 1.5 which, given our sample sizes, indicate that the power of the test at ¼ 0.05 is approximately 50% (Smith et al., 1982). In two of these cases, the deciduous tooth has a greater FA component than the permanent replacement. Likewise, four comparisons yielded F-values between 1.10–1.28 which, given the sample sizes, should have a power of 25% or less. In these comparisons, three permanent teeth have 432 D. GUATELLI-STEINBERG ET AL. TABLE 10. F tests of i2 between groups1 Deciduous vs. male perm. Deciduous vs. female perm. Male vs. female perm. Tooth F value i2 greater in F value i2 greater in F value XC or xc XP1 or xm1 XP2 or xm2 DC or dc DP1 or dm1 DP2 or dm2 1.43 1.12 1.01 1.49 1.10 1.53 Deciduous Deciduous Male perm. Deciduous Male perm. Male perm. 1.56 2.05 1.16 1.28 1.02 1.97 Deciduous Deciduous Female perm. Female perm. Female perm. Female perm. 1.08 1.84 1.14 1.91 1.08 1.28 i2 greater in Male Male Female Female Male Female 1 X, permanent maxillary; D, permanent mandibular; x, deciduous maxillary; d, deciduous mandibular; perm. ¼ permanent. F values in boldface are statistically signiﬁcant. greater FA components than their deciduous predecessors. If the power of the tests was increased so that F-values as low as 1.10 could be conﬁdently accepted as showing statistically signiﬁcant differences, then in this sample there would be ﬁve deciduous teeth and ﬁve permanent teeth with greater FA components. The results of the F-tests are not consistent with our ﬁrst hypothesis, that deciduous teeth should exhibit less FA than their permanent successors, and we reject this hypothesis for this sample. The second hypothesis, that females would have permanent canines with signiﬁcantly less FA than those of males, is also rejected. For the mandibular canine, females show statistically signiﬁcantly greater FA than males. For the maxillary canine, males and females do not differ signiﬁcantly in FA, though FA in the male maxillary canine is higher. The F-value for this comparison is only 1.08 (Table 10). For comparative purposes, data from late prehistoric (LP) and Late Archaic (LA) Ohio Valley Native Americans are included in Figure 1. Note the generally higher i2 values of the Gullah. DISCUSSION This investigation was undertaken to determine if deciduous teeth would exhibit less ﬂuctuating asymmetry than permanent teeth, and if female canines would exhibit less ﬂuctuating asymmetry than those of males. As explained in the introduction, these hypotheses were based on the biology of the developing dentition. Our data, derived from a single Gullah population, reject both hypotheses. We note, however, that the Gullah population we sampled has much higher FA than our samples of prehistoric Ohio Native Americans. Thus, we suggest that the impact on FA of permanent vs. deciduous and male vs. female differences in tooth formation time spans and developmental buffering (intrauterine environment, female development) may be minor when compared to the effect of developmental noise when stress is high. The fact that the Gullah dentition shows generally greater FA than Archaic and late prehistoric Ohio Valley Native Americans is consistent with what is known about population stress from historical descriptions (Gullah) and archaeological evidence (Ohio Valley Indians). There is strong evidence indicating that the Gullah, despite being a 20th century American population, were subjected to high levels of stress throughout their lifetimes. Through at least the early 1960s, economic and social conditions kept the Gullah in poverty, landless, and in jobs requiring hard manual labor (Newby, 1973a). The medical community of South Carolina at the time was aware of the health problems that resulted from these conditions, and several reports are available documenting the Gullah’s (and other African American South Carolinians’) plight. Aycock (1964) found a high correlation between infant deaths and poverty, illiteracy, illegitimacy, and race. In 1951–1952, four times as many blacks as whites used public child health facilities, while the black population of South Carolina was dropping from two thirds to one third of the total population (Aycock, 1964). More than 4% of black children seen in these clinics were malnourished; 0.7% had rickets (Newby, 1973b). In 1960, the overall life expectancy for South Carolina was 66.41 years, the lowest for the country. However, life expectancy in the state was approximately 10 years shorter for black South Carolinians than for white (US Department of Health, Education, and Welfare, 1961). By the 1950s, diseases that had been problematic, such as malaria, hookworm, pellagra, typhoid, smallpox, and diphtheria, could only be found among the poorest South Carolina blacks, including the Gullah. However, venereal disease and tuberculosis were still problematic. During the 1950s and 1960s, approximately 85% of all cases of venereal disease treated in public health clinics were in African Americans (Newby, 1973a). Congenital syphilis and other venereal diseases would likely be a source of developmental stress and possibly dental asymmetry. Skeletal and dental stress indicators in the Ohio Valley Native American populations suggest a ‘‘low prevalence of pathological conditions’’ (Sciulli, 2002, p. 42). Steckel et al. (2002) compared Ohio Valley Native American Archaic and Pearson archaeological samples with other Western Hemisphere archaeological samples in terms of a health index based on stature, enamel hypoplasias, skeletal evidence of anemia, dental pathology, periostitis, trauma, and degenerative joint disease. The health indices in these archaeological populations range from a low of 53.5 (indicating poor health) to 91.8 (indicating good health). The Pearson population health index is 73.7, and that of the Archaic population is 77.9, both above the median score of 72.8. African American slave populations had lower health indices than these, with an average health index of 67.1 (Steckel et al., 2002). The presence of some directional asymmetry in our sample may also reﬂect population-level stress. Harris (1992) suggested that directional asymmetry may be related to developmental timing differences between antimeres, which may in turn result from developmental stress in a population. Sharma et al. (1986) proposed that directional asymmetry might result from environmental or genetic stressors that cause unilateral acceleration of mitotic rates in enamel organs. However, as Townsend and Farmer (1998, p. 253) noted, while directional asymmetry is not a statistical artifact, ‘‘its exact nature and causes remain to be solved.’’ While our data rejected the hypotheses (and we suggest that the high level of population stress in the Gullah is one possible reason for this), there are three other points to consider in interpreting our results. First, although we GULLAH FLUCTUATING DENTAL ASYMMETRY compared deciduous and permanent teeth from a single population, a more reﬁned test of our hypothesis would be to test FA in the deciduous and permanent teeth of the same individuals. The dental casts of the Burlington Growth Study, for example, would provide such an opportunity. Indeed, in the Burlington sample, Saunders and Mayhall (1982) found that deciduous second molars were less asymmetric than permanent molars for the Carabelli trait and protostylid. Interestingly, our data show that maxillary and mandibular m2s have less FA than the other deciduous teeth we measured (maxillary and mandibular canines, and m1s). Liversidge and Molleson (1999) found m2 to be the least variable and least asymmetric tooth of the deciduous dentition, arguing that it may be the polar tooth of the deciduous molar ﬁeld. A second point to consider in interpreting our results surrounds assumptions embedded within our hypothesis that male canines should show more FA than female canines. Although male permanent canines take longer to mineralize than those of females (Moss, 1976; Moss and Moss-Salentijn, 1974), it is not known if the former actually spend more time in the soft-tissue stage of development than the latter. In addition, some cultures may favor children of one sex over another, leading to sex-biased child-care practices that can impact tooth development (e.g., Goodman et al., 1987). While sex certainly played a role in educational and economic opportunities among 20th century South Carolina African Americans, it is not clear from ethnographic records of the Gullah if there were sex biases in child care that may have had health effects (Weiner, 1998). Thirdly, Palmer and Strobeck (2003) noted that FA might be affected by random phenotypic variation that does not result from developmental instability. It is important to note, however, that this factor most affects tissues, such as bone, that have the ability to remodel and are thus phenotypically plastic. Once crown formation is completed, teeth are minimally phenotypically plastic (secondary dentin is deposited, but this does not affect tooth size). In addition, any random phenotypic variation unrelated to developmental instability accruing during the period of crown formation would be expected to involve both sides. The fact that our data reject the hypothesis that male permanent canines are less developmentally stable than those of females is interesting in a comparative nonhuman primate context. Manning and Chamberlain (1993) demonstrated an association across 21 species of Old World primates between male, but not female, FA and canine sexual dimorphism. They interpreted this result to mean that there is a causal link between sexual selection on the canines of nonhuman primate males and ﬂuctuating asymmetry in these teeth. It is thought that sexual selection results in diminished developmental canalization, such that structures are more affected by external conditions affecting growth (Manning and Chamberlain, 1993). The fact that male canines do not exhibit more FA than female canines in our study is consistent with the small degree of sexual dimorphism that human canines exhibit (Kieser, 1990), indicating that human male canines are not undergoing sexual selection for large canine size. In terms of avenues for further investigation, we suggest that future studies aim to 1) compare deciduous teeth and permanent teeth from the same individuals, 2) test these hypotheses in populations which are minimally stressed, and 3) test these hypotheses in populations with no evidence of sex-bias in child care. Basic information regarding potential differences in male-female soft-tissue formation stages is also needed. Finally, we point out that 433 more studies focusing on the issue of developmental stability in the human dentition will ultimately help improve our understanding of ﬂuctuating dental asymmetry as an environmental stress indicator. CONCLUSIONS This study tested two related hypotheses about ﬂuctuating asymmetry (FA): 1) deciduous teeth were predicted to be more developmentally stable than their permanent successors; and 2) female canines were predicted to be more developmentally stable than male canines. To test these hypotheses, bucco-lingual measurements were made on dental casts from a single Gullah population on sample sizes ranging from 63–82 antimeric pairs, using a repeated-measures study design to partition out measurement error from between-sides variance. Neither hypothesis was supported by our data. In most cases, Gullah deciduous teeth did not exhibit statistically signiﬁcantly less FA than their permanent successors; indeed, statistically signiﬁcant differences were found for only 3 of 12 deciduous vs. permanent contrasts, and in two cases, the deciduous tooth had greater FA. Female mandibular canines exhibited statistically signiﬁcantly greater FA than those of males, while there was no statistically signiﬁcant sex difference in FA for the maxillary canine. Compared to Archaic and late prehistoric Ohio Valley Native Americans, FA in these Gullah samples is high. This FA difference is consistent with historical and archaeological evidence that environmental stress was higher in the Gullah population relative to these Native American groups. LITERATURE CITED Aycock EK. 1964. Infant mortality in South Carolina. J SC Med Assoc 60:1–5. Bailit HL, Workman PL, Niswander JD, MacLean CJ. 1970. Dental asymmetry as an indicator of genetic and environmental conditions on human populations. Hum Biol 42:626–638. Cook DC, Buikstra JE. 1979. Health and differential survival in prehistoric populations: prenatal dental defects. Am J Phys Anthropol 51:649–664. D’Agostino RB. 1986. Tests for the normal distribution. In: D’Agostino RB, Stephens MA, editors. Goodness of ﬁt techniques. New York: Marcel Dekker. p 367–419. Falcolner DS, Mackay TFC. 1997. Introduction to quantitative genetics, 4th ed. Essex: Longman. Fraser FC. 1994. Developmental instability and ﬂuctuating asymmetry in man. In: Markow TA, editor. Developmental instability: its origins and evolutionary implications. Boston, MA: Kluwer Academic. p 319–334. Garn SM, Lewis, AB, Kerewsky RS. 1965. Genetic, nutritional, and maturational correlates of dental development. J Dent Res 44:228–242. Garn SM, Lewis AB, Kerewsky RS. 1966. The meaning of bilateral asymmetry in the permanent dentition. Angle Orthod 36:55–62. Goodman AH, Allen LH, Hernandez GP, Amador A, Arriola LV, Chavez A, Pelto GH. 1987. Prevalence and age at development of enamel hypoplasias in Mexican children. Am J Phys Anthropol 72:7–19. Harris EF. 1992. Laterality in human odontometrics: analysis of a contemporary American white series. In: Lukacs JR, editor. Culture, ecology, and dental anthropology. Delhi: Kamla-Raj. p 157–170. Harris EF. 2002. Dental development and anomalies in craniosynostoses and facial clefting. In: Mooney MP, Siegel MI, editors. Understanding craniofacial anomalies: the etiopathogenesis of craniosynostoses and facial clefting. New York: WileyLiss. p 425–467. 434 D. GUATELLI-STEINBERG ET AL. Harris EF, Nweeia MT. 1980. Dental asymmetry as a measure of environmental stress in Ticuna Indians of Colombia. Am J Phys Anthropol 53:133–142. Hershokovitz I, Livshits G, Moskona D, Arensburg B, Kobyliansky E. 1993. Variables affecting dental ﬂuctuating asymmetry in human isolates. Am J Phys Anthropol 91:349–365. Kieser JA. 1990. Human adult odontometrics: the study of variation in adult tooth size. Cambridge: Cambridge University Press. Kieser JA. 1992. Fluctuating odontometric asymmetry and maternal alcohol consumption. Ann Hum Biol 19:513–520. Kieser JA, Groeneveld HT, Preston CB. 1986. Fluctuating dental asymmetry as a measure of odontogenetic canalization in man. Am J Phys Anthropol 71:437–444. Kieser JA, Groeneveld HT, Da Silva PCF. 1997. Dental asymmetry, maternal obesity, and smoking. Am J Phys Anthropol 102:133–139. Lens L, Van Dongen S, Kark S, Matthysen E. 2002. Fluctuating asymmetry as an indicator of ﬁtness: can we bridge the gap between studies? Biol Rev 77:27–38. Liversidge HM, Molleson TI. 1999. Deciduous tooth size and morphogenetic ﬁelds in children from Christ Church, Spitalﬁelds. Arch Oral Biol 44:7–13. Manning JT, Chamberlain AT. 1993. Fluctuating asymmetry, sexual selection and canine teeth in primates. Proc R Soc Lond B 251:83–87. Menegaz-Bock RM. 1968. An investigation of the genetic basis for structural relationship in the anterior dentition. Ph.D. thesis, University of Chicago. Mizoguchi Y. 1980. Factor analysis of environmental variation in the permanent dentition. Bull Natl Sci Mus Tokyo Ser D 6:29–46. Mizoguchi Y. 1998. Size and co-variation among deciduous teeth. In: Lukacs JR, editor. Human dental development, morphology, and pathology: a tribute to Albert A. Dahlberg. University of Oregon anthropological papers number 54. Eugene: University of Oregon. p 259–285. Moss ML. 1978. Analysis of developmental processes possibly related to human dental sexual dimorphism. In: Butler BM, Joysey KM, editors. Development, evolution, and function of teeth. New York: Academic Press. p 135–148. Moss ML, Moss-Salentijn L. 1976. Analysis of developmental processes possibly related to human dental sexual dimorphism in permanent and deciduous canines. Am J Phys Anthropol 46:407–414. Narayanan A, Smith S, Townsend G. 1999. Dental crown size in individuals with cleft lip and palate. In: Townsend G, Kieser J, editors. Perspectives in human biology: dentofacial variation in perspective. Perth: Centre for Human Biology. p 61–70. Newby IA. 1973a. South Carolina Department of Health seventythird annual report 1951–52. Charleston, SC: South Carolina Department of Health. p 308–309. Newby IA. 1973b. Black Carolinians: a history of blacks in South Carolina from 1895 to 1968. Columbia: University of South Carolina Press. Niswander JD, Chung CS. 1965. The effects of inbreeding on tooth size in Japanese children. Am J Hum Genet 17:390–398. Palmer AR. 1994. Fluctuating asymmetry analyses: a primer. In: Markow TA, editor. Developmental instability: its origins and evolutionary implications. Boston: Kluwer Academic. p 335–364. Palmer AR, Strobeck C. 1986. Fluctuating asymmetry: measurement, analysis, patterns. Ann Rev Ecol Syst 17:391–422. Palmer AR, Strobeck C. 2003. Fluctuating asymmetry analyses revisited. In: Polak M, editor. Developmental instability. Oxford: Oxford University Press. p 279–319. Quetisch C, Sonster TA, Lindquist S. 2002. Hsp90 as a capacitor of phenotypic variation. Nature 417:6187–624. Saunders SR, Mayhall JT. 1982. Fluctuating asymmetry of dental morophological traits: new interpretations. Hum Biol 54:789–799. Scheffé H. 1959. The analysis of variance. New York: John Wiley and Sons. Sciulli PW. 1978. Developmental abnormalities of the permanent dentition in prehistoric Ohio Valley Amerindians. Am J Phys Anthropol 48:93–98. Sciulli PW. 2002. Dental asymmetry in a Late Archaic and late prehistoric skeletal sample of the Ohio Valley Area. Dent Anthropol 16:33–44. Sciulli PW, Doyle WJ, Kelly S, Siegel P, Siegel MI. 1979. The interaction of stressors in the induction of increased levels of ﬂuctuating asymmetry in the laboratory rat. Am J Phys Anthropol 50:279–284. Sharma K, Corruccini RS, Potter RHY. 1986. Genetic and environmental inﬂuences on bilateral dental asymmetry in Northwest Indian twins. Int J Anthropol 4:349–360. Siegel MK, Doyle W. 1975a. The differential effects of prenatal and postnatal audiogenic stress on ﬂuctuating dental asymmetry. J Exp Zool 191:211–214. Siegel MK, Doyle W. 1975b. The effects of cold stress on ﬂuctuating asymmetry in the dentition of the mouse. J Exp Zool 193:385–389. Siegel MI, Doyle W, Kelly C. 1977. Heat stress, ﬂuctuating asymmetry and prenatal selection in the laboratory rat. Am J Phys Anthropol 46:121–126. Smith BH, Garn SM, Cole PE. 1982. Problems of sampling and inference in the study of ﬂuctuating dental asymmetry. Am J Phys Anthropol 58:281–289. Steckel RH, Sciulli PW, Rose JC. 2002. A health index for skeletal remains. In: Steckel RH, Rose JC, editors. The backbone of history: health and nutrition in the Western Hemisphere. Cambridge: Cambridge University Press. p 61–93. Stinson S. 1985. Sex differences in environmental sensitivity during growth and development. Yrbk Phys Anthropol 28: 123–147. Townsend GC. 1981. Fluctuating asymmetry in the deciduous dentition of Australian Aboriginals. J Dent Res 60:1849–1857. Townsend GC. 1983. Fluctuating dental asymmetry in Down’s syndrome. Aust Dent J 28:39–44. Townsend GC, Brown T. 1980. Dental asymmetry in Australian Aboriginals. Hum Biol 52:661–673. Townsend GC, Farmer V. 1998. Dental asymmetry in the deciduous dentition of South Australian children. In: Lukacs JR, editor. Human dental development, morphology, and pathology: a tribute to Albert A. Dahlberg. University of Oregon anthropological papers number 54. Eugene, University of Oregon. p 245–257. Townsend GC, Garcia-Godoy F. 1984. Fluctuating asymmetry in the deciduous dentition of Dominican Mulatto children. Arch Oral Biol 29:483–486. US Department of Health, Education, and Welfare. 1961. Public health service, state life tables 1959–61, volume II, no. 41, South Carolina. Washington, DC: US Department of Health, Education, and Welfare. p 568. Van Valen L. 1962. A study of ﬂuctuating asymmetry. Evolution 16:125–142. Weiner MF. 1998. Mistresses and slaves: plantation women in South Carolina, 1830–80. Urbana: University of Illinois Press. Zar JH. 1999. Biostatistical analysis, 4th ed. Upper Saddle River: Prentice Hall.