Biology Reference
In-Depth Information
Williams, 2008; Sholts et al., 2011a
) have evaluated the precision of various landmark types
and error associated with different coordinate observation techniques. In summary, these
studies find a fairly high degree of precision and repeatability for Type I and II landmarks
across different observation modalities. Further, thorough knowledge of landmark defini-
tions and locations and practice using a digitizer increase a researcher's precision when
observing landmark coordinates (yes, practice, practice, practice, and then practice some
more).
When conducting research that includes coordinate data, an assessment of measurement
error should be part of the research design. Currently, there are several methods for evalu-
ating the precision and repeatability of landmark coordinate data. The most popular
approaches are critically reviewed by von Cramon-Taubadel and colleagues (2007) and the
authors propose an alternate method for assessing measurement error. Anyone interested
in using coordinate data should carefully read this article before choosing a method for their
research.
Semilandmarks
Some of the original data types of interest in geometric morphometrics were as outlines or
curves, such as the midsagittal profile of the frontal bone. In humans we know there are
differences in the curvature of the frontal bone between the sexes, across human populations,
and through time due to secular change. While the curve or outline may be anchored by Type
I landmarks at either end, the intervening coordinates that define the curve or outline lack
homology, necessitating different methods to enable meaningful analyses. This is an impor-
tant endeavor as many of the anatomical regions of interest are not defined by Type I or II
landmarks, yet these surfaces and curves contain important information for interpreting
aspects of morphological variation. The ability to explore this variation by quantifying the
curvature and testing differences with rigorous statistical analysis would be beneficial to
studies of human variation.
Semilandmarks (also called sliding landmarks) can be acquired from curves and outlines
by tracing with a three-dimensional digitizer set to collect coordinates spaced equidistant
apart (e.g., every 0.05 mm) and downloaded to a custom program such as 3Skull (described
earlier), which can record coordinates along an arc, or to a utility program such as Microsoft
Excel. Semilandmarks can also be sampled from laser surface scans or from volumetric (CTor
MRI) scans. The custom program, Landmark Editor (see
Table 12.1
), can be used to place
semilandmarks across surfaces captured with laser scans.
Currently, most geometric morphometric research in biological anthropology employs
landmark coordinates observed on the cranium. The long-standing interest in craniofacial
variation, the ease of data collection, the numerous biologically meaningful landmarks on
the cranium, and the suite of available sophisticated analytical packages make this method
a natural inclination. Nevertheless, studies utilizing landmark coordinates characterizing
the morphology of postcranial elements do exist (e.g., Stephens and Strand Viðarsd´ ttir,
2008;
Bytheway and Ross, 2010; De Groote et al., 2010; De Groote, 2011a, 2011b
). It is likely
that similar studies will be more plentiful in the future. Although the methodological devel-
opments regarding the use of semilandmarks are relatively new, there are now several
programs capable of analyzing semilandmarks, and it is expected that this approach will
see significant growth in the coming years.
