close

Вход

Забыли?

вход по аккаунту

?

Comment on УAlternative Dental MeasurementsФ by Hillson et al.

код для вставкиСкачать
AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 132:234?237 (2007)
Notes and Comments
Comment on ??Alternative Dental Measurements?? by Hillson et al.
Christopher M. Stojanowski
Center for Bioarchaeological Research, School of Human
Evolution and Social Change, Arizona State University,
Tempe, Arizona
In a recent AJPA paper, Hillson et al. (2005) proposed a
series of dental measurements as alternatives to the
traditional mesiodistal and buccolingual crown diameters.
Their reasoning was straightforward. Dental attrition is
the most significant cause of missing data in prehistoric
odontometric research, and measurements should be
identified that help mitigate this problem. Other approaches, generally more labor-intensive, have been
proposed (Biggerstaff, 1969, 1975; Corruccini, 1977;
Corruccini and Potter, 1981; Mayhall, 2000; Morris,
1986), including cervical dimensions (Alt et al., 1998;
Corruccini, 1977), reflective of the long-term sensitivity of
odontometrists to this problem (see Goose, 1963). The
continued use of crown dimensions, in particular mesiodistal diameters, necessitates the use of either strictly
univariate approaches or significant data imputation prior
to implementation of most packaged multivariate statistical applications. In my own experience (Stojanowski,
2001, 2005), it is difficult for even the trained observer to
completely cull teeth minimally affected by approximal
attrition; therefore, age-dependency remains a significant
concern. Hillson et al. (2005) provided five contributions to
odontometry: 1) they clearly defined mesiodistal and
buccolingal cervical dimensions, 2) they introduced a
caliper designed to accurately record these dimensions,
3) they defined maximum diagonal molar dimensions, 4)
they calculated correlations between these new measurements and traditional crown diameters, and 5) they tested
the reliability (observer error) of the new dimensions.
Using a series of dental remains from the 14th century
Smith?eld site, Hillson et al. (2005) demonstrated that
these alternative dental dimensions were as reliably
recorded as crown diameters. Tooth-speci?c crown-cervical buccolingual and mesiodistal correlations were moderate in magnitude, averaging 0.775 for the former and
0.634 for the latter. This suggests that crown and cervical dimensions re?ect similar, but not identical, tooth
components (whether genetic, environmental, or both).
Correlations between within-tooth mesiodistal vs. buccolingual dimensions were slightly lower, averaging 0.599
for the crown and 0.569 for cervical dimensions. Similarities indicate that tooth-shape variation (length by width)
was roughly similar between measurement sets.
Although the results of Hillson et al. (2005) are promising, two issues remain uncertain and are tested in this
paper: 1) What are the actual sample-size bene?ts of cervicometrics, and which teeth are most likely to bene?t
from substitution? 2) Do crown and cervical dimensions
re?ect similar aspects of the dentition? That is, because
odontometrics are used in a variety of applications (biological af?nity, sexual dimorphism, and directional or
C 2006
V
WILEY-LISS, INC.
?uctuating asymmetry; see Colby, 1996; Corruccini,
1977; Falk and Corruccini, 1982; Kieser, 1990), it is
unclear whether cervicometrics produce equivalent
results (for a similar query, see Alt et al., 1998).
To evaluate whether cervical dimensions can be used
as proxies for homologous crown metrics, data were collected from the Windover Pond site, an early Archaic period (ca. 7,500 years BP) mortuary pond on the Atlantic
coast of Florida (Doran, 2002). The following statistics
were enumerated:
1. Quantify sample size bene?ts of cervical dimensions,
and identify those teeth most likely to bene?t from
substitution.
2. Compare the correlation results of Hillson et al.
(2005) to a different (geographically and temporally)
population group.
3. Determine whether cervical dimensions demonstrate
more or less variability than crown dimensions.
4. Compare side asymmetry to determine if the pattern
and magnitude of asymmetry differed between crown
and cervical dimensions. This also speaks to the
appropriateness of antimere substitution.
5. Compare interindividual distances, and determine if
crown and cervical data re?ect similar biological relationships (see also Falk and Corruccini, 1982).
Dental data were recorded following the conventions
of Moorrees and Reed (1954) and Hillson et al. (2005) for
crown and cervical dimensions, respectively. Mesiodistal
and buccolingual diameters were recorded for right and
left dentitions to the nearest 0.10 mm, using HillsonFitzgerald calipers. Visibly worn teeth were not included
in this analysis. Simple inferential hypothesis tests and
Pearson correlation coef?cients were used to assess the
correspondence among measurement sets. Principal components analysis was used to assess multivariate relationships among variables and between a subset of the
best-preserved burials.
RESULTS
Data on the representation of mesiodistal and buccolingual diameters are presented in Table 1. Enumerated statistics represent the number of teeth for which only crown
or cervical dimensions were recorded, the difference
between these statistics, and the combined (left and right)
increase in sample size and percent improvement (or not)
offered by the use of cervical dimensions. Results for mesiodistal dimensions were encouraging. Although sample sizes
*Correspondence to: Christopher M. Stojanowski, Center for Bioarchaeological Research, School of Human Evolution and Social
Change, Arizona State University, Tempe, AZ 85287.
Received 8 September 2005; accepted 16 November 2005.
DOI 10.1002/ajpa.20400
Published online 9 May 2006 in Wiley InterScience
(www.interscience.wiley.com).
235
NOTES AND COMMENTS
TABLE 1. Sample sizes for crown and cervical dimensions
Left
Cr
Cer
Mesiodistal
6 17
I1
I2
3 13
C
7 23
1
2 22
P
2 29
P2
4 33
M1
2
8 19
M
0 16
I1
2 19
I2
C
5 27
3 23
P1
2 22
P2
4 38
M1
9 28
M2
Buccolingual
1
8
4
I
2
3
I2
C
3
8
3
3
P1
2
3 13
P
6 10
M1
5
M2 15
4
3
I1
8
3
I2
C
9 11
3
2
P1
4
8
P2
4
9
M1
6
M2 13
Right
Diff
Cr
Cer
Diff
Combined
TABLE 2. Descriptive statistics, F-tests, and
correlations: left side
%
Percentage
�
�
�
�
�
�
�
�
�
�
�
�
�
�
4
10
7
4
1
3
7
2
2
9
4
2
1
5
13
12
25
27
29
35
19
21
25
25
26
26
39
24
���
�
�
�
�
�
�
�
�
�
�
�
�
�
�
�
�
�
�
�
�
�
�
�
�
�
14%
9%
24%
31%
39%
44%
16%
25%
29%
27%
30%
31%
51%
27%
4
��0
�
�10
1
5
�1
��7
11
4
5
4
5
9
15
5
4
4
4
4
5
5
2
2
10
10
13
10
5
6
1
9
4
6
13
9
9
2
����10
�3
�0
���
13
1
�
��
�20
0
8
�1
��
3
9%
0%
7%
4%
13%
4%
14%
0%
6%
5%
0%
4%
9%
2%
Crown
I1
2
I
C
P1
P2
M1
M2
I1
I2
C
P1
P2
M1
M2
Note: Cr � crown; Cer � cervical; Diff � difference between
crown and cervical sample sizes.
in reference to the theoretical maximum (n � 140) were
low for all measurements, cervical dimensions were better
represented for all teeth and both sides. The largest percent increase was observed for maxillary and mandibular
?rst molars. The overall average increase in sample size
for all teeth was 28%. For buccolingual dimensions, the
results were mixed. Although the overall average increase
in representation was positive (1%), several teeth were
actually cervically less well-represented in comparison to
homologous crown dimensions. The lack of sample-size
bene?t results from the similar locations of both measurement types, and the fact that buccolingual dimensions, apically displaced in comparison to mesiodistal dimensions,
are less affected by attrition at an earlier age as a result.
Crown and cervical dimensions did not differ in patterns of phenotypic variation (Table 2). Although several
individual F-tests were signi?cant (those critical values
greater than 2), these differences were not robust to family-wise error adjustments. For those marginal cases, however, there was no preference as to which measurement
type was more variable, suggesting no systematic differences in tendency toward ontogenetic plasticity. These data
suggest that cervical dimensions demonstrate similar levels of variability in comparison to homologous crown
measurements, and may, therefore, serve as appropriate
proxies in analyses of comparative phenotypic variation.
Whereas Hillson et al. (2005) found all positive correlations between homologous cervical and crown diameters with slightly higher anterior tooth correlations,
Windover produced a different pattern (Table 2). Interpreting data from the maxilla was dif?cult due to sample-size limitation and the effect this had on statistical
MD
BL
MD
BL
MD
BL
MD
BL
MD
BL
MD
BL
MD
BL
MD
BL
MD
BL
MD
BL
MD
BL
MD
BL
MD
BL
MD
BL
Cervical
N
SD
N
SD
F
r
L-R
(Cr)
L-R
(Cerv)
9
26
12
23
14
29
15
29
13
26
21
38
21
45
10
22
13
27
12
28
10
24
9
26
13
34
15
35
0.37
0.54
0.95
0.56
0.43
0.87
0.40
0.62
0.46
0.59
0.61
0.54
0.57
0.64
0.35
0.45
0.53
0.29
0.44
0.55
0.42
0.49
0.62
0.46
0.79
0.54
0.82
0.62
20
22
22
23
33
34
36
29
39
36
49
42
32
36
26
22
29
22
34
30
30
23
29
30
47
39
34
28
0.69
0.39
0.73
0.52
0.75
0.54
0.44
0.55
0.38
0.83
0.79
0.59
0.53
0.84
0.27
0.27
0.28
0.25
0.53
0.54
0.30
0.35
0.43
0.56
0.57
0.48
0.67
0.67
3.47
1.92
1.69
1.16
3.04
2.60
1.21
1.27
1.47
1.98
1.68
1.19
1.16
1.72
1.68
2.78
3.58
1.35
1.45
1.04
1.96
1.96
2.08
1.48
1.92
1.27
1.50
1.68
0.86
0.54*
0.02
0.87*
0.20
0.70*
0.31
0.73*
0.35
0.86*
0.62*
0.66*
0.29
0.68*
0.06
0.82*
0.03
0.61*
0.43
0.85*
0.44
0.71*
0.30
0.44*
0.55
0.51*
0.53
0.44*
0.82*
0.90*
0.89*
0.81*
0.84*
0.89*
0.84*
0.91*
0.73*
0.92*
0.78*
0.66*
0.47
0.60*
0.73
0.84*
0.64
0.71*
0.91*
0.82*
0.72*
0.82*
0.81*
0.83*
0.22
0.87*
0.71
0.84*
0.21
0.77
0.23
0.85*
0.20
0.94*
0.44
0.90*
0.54
0.78*
0.67*
0.51*
0.71*
0.54*
0.64*
0.90*
0.70*
0.58*
0.94*
0.88*
0.79*
0.84*
0.85*
0.82*
0.54*
0.79*
0.65
0.83
*Signi?cant at 5% level. P-values were adjusted for family-wise
error, using individual a/n � 0.001 to maintain overall 5% error rate.
Note: MD � mesiodistal; BL � buccolingual.
power. In general, mesiodistal correlations were not signi?cant, because very few individuals had both crown
and cervical measurements observable.1 Buccolingual
dimensions, however, were all signi?cantly correlated in
a manner similar to the results of Hillson et al. (2005).
Anterior teeth produced larger within-tooth correlations
than molars, although most Windover correlations were
generally lower than those previously reported. For the
mandible, mesiodistal correlations were lower than
within-tooth buccolingual dimensions, and all were not
signi?cant. While Hillson et al. (2005) found a similar
pattern, several measurements in the present study
returned correlations very close to 0. Mandibular anterior buccolingual dimensions were clearly more strongly
correlated than molar dimensions. Interestingly, four
cervical dimensions, all at P1 or anterior, were negatively correlated with homologous crown measurements.
All four were mesiodistal diameters. Whereas P1 demonstrated patterns more typical of anterior teeth, P2 was
more similar to molars in the pattern of within-tooth
crown-cervical correlations.
Principal components analysis of mesiodistal and buccolingual cervical dimensions (Table 3) returned seven
components with eigenvalues greater than 1, explaining
81% of the variation in the original data set. Unlike
many principal components analyses of phenotypic variation in which principal component (PC) 1 demonstrates
1
In general, individuals with unworn mesiodistal anterior dimensions were too young for cervical completion, and individuals with
complete cervico-enamel junctions had already been affected by
interproximal attrition.
American Journal of Physical Anthropology?DOI 10.1002/ajpa
236
NOTES AND COMMENTS
TABLE 3. Principal components analysis of mesiodistal and
buccolingual cervical diameters and buccolingual crown and
cervical diameters
Cervical
Measure
1
PC1
Crown and cervical
PC2
BLI
MDI2
BLI2
MDC
BLC
MDP1
BLP1
MDP2
BLP2
MDM1
BLM1
MDM2
BLM2
MDI1
BLI1
MDI2
BLI2
BLC
MDP1
BLP1
BLP2
MDM1
BLM1
MDM2
BLM2
0.465
0.127
0.546
0.214
0.669
0.010
0.651
0.184
0.425
0.768
0.768
0.685
0.623
0.442
0.522
0.373
0.370
0.676
0.169
0.376
0.756
0.703
0.706
0.739
0.683
0.340
0.472
0.053
0.140
0.166
0.016
0.388
0.270
0.444
0.030
0.190
0.499
0.465
0.482
0.328
0.536
0.521
0.338
0.101
0.554
0.410
0.395
0.001
0.076
0.232
Eigenvalue
Variance %
7.648
30.59
3.026
12.10
Measure
1
CBLI
NBLI1
CBLI2
NBLI2
CBLC
NBLC
CBLP1
NBLP1
CBLP2
NBLP2
CBLM1
NBLM1
CBLM2
NBLM2
CBLI1
NBLI1
CBLI2
NBLI2
CBLC
NBLC
CBLP1
NBLP1
CBLP2
NBLP2
CBLM1
NBLM1
PC1
PC2
0.693
0.541
0.708
0.624
0.751
0.601
0.717
0.721
0.777
0.516
0.561
0.814
0.740
0.600
0.611
0.506
0.108
0.258
0.694
0.469
0.737
0.668
0.554
0.655
0.726
0.753
10.608
40.81
0.081
0.154
0.161
0.104
0.232
0.148
0.174
0.427
0.310
0.426
0.116
0.079
0.202
0.239
0.461
0.592
0.744
0.646
0.014
0.064
0.010
0.232
0.548
0.510
0.160
0.043
2.931
11.27
large positive loadings for all variables (representing a
general size factor), cervical metrics display a more complicated pattern. While all mandibular measurements,
and maxillary molar measurements, load positively
along PC1, maxillary anterior mesiodistal dimensions
demonstrate small positive or even negative loadings.
Therefore, PC1 represents cervical size exclusive of the
mesiodistal anterior maxillary dentition. PC2, while
more dif?cult to interpret, appears to model shape variation related to anterior mandibular dimensions (positive)
vs. posterior maxillary dimensions (negative). PC3?7
loadings were not interpretable.
Consideration of the relationship among buccolingual
crown and cervical dimensions produced a more typical
pattern (Table 3). Five components returned eigenvalues
greater than 1, representing 75% of the variation in the
original data set. PC1 exhibited positive loadings for all
variables, re?ecting a size component of the dentition,
while PC2 appeared to represent shape differences
related to mandibular incisor size (positive) vs. maxillary
and mandibular premolar size (negative). That components did not segregate crown and cervical dimensions
suggests similar morphological etiologies, and is consistent with the correlation data presented above. PC3?5
loadings were not interpretable.
Asymmetry data are also presented in Table 2. Leftright crown correlations were moderate to large in magnitude, and for the most part signi?cantly different from
0. Molar side correlations tended to be lower than anterior tooth dimensions. The lowest crown side correlation
was for mesiodistal M1, and the mandibular dimensions
as a whole (0.74) were less well-correlated between sides
than maxillary diameters (0.79). Cervical dimensions,
however, demonstrated a slightly different pattern. While
all mandibular correlations, with the exception of mesiodis-
tal and buccolingual M2, were moderate to large and signi?cantly different from 0, and were actually better correlated between sides than mandibular crown dimensions
(0.76 vs. 0.74), the maxillary cervical data produced an
interesting pattern. Side correlations for P2 through M2
were moderate in magnitude, comparable to homologous
crown correlations, and signi?cantly different from 0. For P1
and mesial teeth, however, buccolingual diameters returned
moderate to large correlations that were signi?cantly different from 0, while maxillary anterior mesiodistal dimensions
were poorly correlated between sides and not signi?cantly
different from 0. Overall, maxillary cervical dimensions
demonstrated the most evidence for asymmetry (average
left-right correlation of 0.59).
Finally, I compared interindividual af?nity represented
by homologous crown and cervical dimensions to determine
whether similar patterns of biological relationships are
represented. Two multivariate plots of principal component loadings generated from the complete series of buccolingual measurements are presented in Figure 1. Both
plots display similar relationships, somewhat visually
obscured by the differential placement of individual 3,
who is centrally located in Figure 1b but more peripheral in Figure 1a. If moved to a more positive position
along PC2 (Fig. 1b), crown and cervical data would produce nearly identical relationships: individuals 6 and 7
to the extreme left, 2 at bottom right, 3 and 5 at upper
right, and 1, 4, and 8 in the center. Overall, these data
are encouraging and suggest similar topologies for crown
and cervical metrics. Minor differences exist when the
scale of differentiation is minimized and closer to the
center of the sampling distribution.
CONCLUSIONS
Cervical dimensions are useful addenda to the standard odontometric data collection strategy, adding further
support to the conclusions of Hillson et al. (2005). Several caveats are offered here, based on data collected
from Windover Pond:
1. Results of comparative preservation analyses produced mixed results, re?ecting the fact that cervicometrics are also subject to multiple causes of missing
data. These include: extensive attrition and subsequent root-splitting, root loss and cervical abrasion
with cingulum chipping, cervical caries, inter-proximal grooves, idiosyncratic use-related attrition, malocclusion, and most signi?cantly, calculus. Mesiodistal
cervical dimensions, often distant in location from homologous crown measures, were much better represented at Windover. However, buccolingual cervical
dimensions were not better represented, and for some
measurements were actually less likely to be present
than homologous crown dimensions.
2. The sparse buccolingual data matrix is unfortunate,
because buccolingual cervicometrics would serve as
an appropriate proxy for crown dimensions. This is
evidenced by:
American Journal of Physical Anthropology?DOI 10.1002/ajpa
a. Similar degrees of phenotypic variation.
b. Moderate to large crown-cervical correlation coef?cients that were statistically signi?cant.
c. Similar PC loadings and loading patterns, including eigenvectors that did not divide variables into
crown and cervical subsets.
d. Similar patterns of dental asymmetry.
e. Similar patterns of interindividual biological af?nity.
237
NOTES AND COMMENTS
measurements are not loading for general dental
size.
c. Small left-right side correlations for anterior maxillary cervicometrics.
In conclusion, buccolingual cervical dimensions appear to
represent similar aspects of human dental morphology and
could serve as useful proxies, despite the limited samplesize bene?ts offered at Windover. Mesiodistal cervical
dimensions, and particularly anterior maxillary measurements, were not suitable proxies for homologous crown
metrics at Windover. These variables may represent different components of dental morphology, which, based on the
results presented here, appear more susceptible to environmental perturbation than other dental measurements. It is
unclear why the disparity exists between Windover and the
Smith?eld site. It may be signi?cant to note the excessive
maxillary anterior malocclusion and crowding witnessed at
Windover. Ultimately, heritability studies are needed on
larger data sets to further de?ne the complex interactions
documented in the anterior dentition. Nevertheless, cervicometrics should certainly be incorporated into standard
osteological data collection procedures in the future.
LITERATURE CITED
Fig. 1. Principal components loading plots of eight individuals selected from Windover Pond, based on preservation of odontometric data. Note similar positions of individuals, with exception of individual 3. a: Crown dimensions. b: Cervical dimensions.
3. Anterior maxillary mesiodistal cervical dimensions
were not, however, an appropriate proxy for homologous crown dimensions. This is evidenced by:
a. Low correlations between cervical and crown anterior (here, P1 and mesial) maxillary and mandibular dimensions.
b. PC1 loadings (size) that did not include maxillary
anterior tooth dimensions, indicating that these
Alt KW, Riemensperger B, Vach W, Krekeler G. 1998. Zahnwurzella?nge und Zahnhalsdurchmesser als Indikatoren zur Geschlechtsbestimmung an menschlichen Za?hnen. Anthropol Anz
56:131?144.
Biggerstaff RH. 1969. The basal area of posterior tooth crown
components: the assessment of within tooth variation of premolars and molars. Am J Phys Anthropol 31:163?170.
Biggerstaff RH. 1975. Cusp size, sexual dimorphism, and heritability of cusp size in twins. Am J Phys Anthropol 42:127?139.
Colby GR. 1996. Analysis of dental sexual dimorphism in two
western Gulf of Mexico precontact populations utilizing cervical measurements. Am J Phys Anthropol 22:87.
Corruccini RS. 1977. Crown component variation in hominoid
lower third molars. Z Morphol Anthropol 68:14?25.
Corruccini RS, Potter RH. 1981. Developmental correlates of
crown component asymmetry and occlusal discrepancy. Am J
Phys Anthropol 55:21?31.
Doran GH. 2002. Windover: multidisciplinary investigations of an
early Archaic period Florida cemetery. Gainesville: University
Press of Florida.
Falk D, Corruccini R. 1982. Ef?cacy of cranial versus dental
measurements for separating human populations. Am J Phys
Anthropol 57:123?128.
Goose DH. 1963. Dental measurement: an assessment of its
value in anthropological studies. In: Brothwell DR, editor.
Dental anthropology. New York: Pergamon. p 125?148.
Hillson S, FitzGerald C, Flinn H. 2005. Alternative dental
measurements: proposals and relationships with other measurements. Am J Phys Anthropol 126:413?426.
Kieser JA. 1990. Human adult odontometrics. Cambridge: Cambridge University Press.
Mayhall JT. 2000. Dental morphology: techniques and strategies.
In: Katzenberg MA, Saunders SR, editors. Biological anthropology of the human skeleton. New York: Wiley-Liss. p 103?134.
Moorrees CFA, Reed RB. 1954. Correlations among crown diameters of human teeth. Arch Oral Biol 9:685?697.
Morris DH. 1986. Maxillary molar occlusal polygons in ?ve
human samples. Am J Phys Anthropol 70:333?338.
Stojanowski CM. 2001. Cemetery structure, population aggregation and biological variability in the mission centers of La
Florida. Ph.D. dissertation, Department of Anthropology, University of New Mexico.
Stojanowski CM. 2005. Biocultural histories in La Florida: a
bioarchaeological perspective. Tuscaloosa: University of Alabama Press.
American Journal of Physical Anthropology?DOI 10.1002/ajpa
Документ
Категория
Без категории
Просмотров
4
Размер файла
128 Кб
Теги
measurements, hillson, dental, comments, уalternative
1/--страниц
Пожаловаться на содержимое документа