Biology Reference
In-Depth Information
both shape and size, geometric morphometric methods provide significant insight into how
each contributes to dimorphism and clarify the shape differences between males and females
that exist independent of size.
Using a generalized Procrustes analysis (GPA) of three-dimensional coordinates from 16
craniofacial landmarks observed on Euro-American and African-American males and
females of known sex,
Kimmerle et al. (2008)
investigated shape and size variation associated
with sexual dimorphism. A MANCOVA procedure employing principal component scores
derived from the Procrustes coordinates and Centroid Size indicated that sex had a significant
effect on shape for both groups, but that size did not influence shape for either group
(meaning that within each sex, shape was the same regardless of size). Based on this finding,
they suggest that varying degrees of sexual dimorphism across human populations may be
the result of size differences, rather than allometric variation.
Gonzalez et al. (2011)
used crania from the Coimbra collection in Portugal to explore
patterns of sexually dimorphic shape and size in the cranium. They used 12 landmarks and
25 semilandmarks to quantify the shape of the glabella and the malar, mastoid, and frontal
processes. The PCA results indicated a low degree of dimorphism in shape and discriminant
analysis produced poor classification results. When Centroid Size was included in the
analysis, correct classification increased, reflecting the marked difference in size between
the sexes.
Sexual dimorphism in subadults was investigated by
Franklin et al. (2007)
utilizing
a sample of 96 known subadult mandibles using coordinates from 38 landmarks. Results
indicated little sexual dimorphism in the subadult mandible for all populations and classifi-
cation results were poor (only 59% correctly classified overall). However, the authors found
some significant shape variables associated with different populations. The authors suggest
that variation in mandibular shape may be attributed to morphology established early in
ontogeny and the result of inherited genetic traits.
In an effort to quantify morphological traits typically assessed through visual observation
for estimating sex,
Pretorius et al. (2006)
analyzed a sample of South Africans comparing
three areas that may be sexually dimorphic: the shape of the greater sciatic notch, mandib-
ular ramus flexure, and shape of the orbits. The coordinate data were analyzed using rela-
tive warp scores, thin-plate splines, and CVA. As expected, the greater sciatic notch
provided the best separation of the sexes. Surprisingly, a greater degree of accuracy for esti-
mating sex was obtained from the orbits as opposed to the ramus.
Bigoni et al. (2010)
also
found a high degree of classification using landmarks of the orbits, suggesting that the orbits
are a region of sexual dimorphism that previously was not identified using traditional
methods.
In application of geometric morphometric methods to postcranial elements,
Bytheway and
Ross (2010)
attempted to define landmarks on the os coxa that might be useful for quantifying
the morphological variation employed in sex estimation. Three-dimensional coordinates for 36
landmarks were observed on adult os coxae from a sample composed of African-Americans
and Euro-Americans. GPA was used to fit the coordinates, which were then subjected to
PCA. The PC scores were utilized as data for a MANCOVA and discriminant analysis. Results
indicated that both size and sex have a significant effect for both groups and that these land-
marks produce highly accurate sex estimation. For more information on sexual dimorphism
and sex estimation/assessment, see Moore (Chapter 4), this volume.
