Biology Reference
In-Depth Information
CHAPTER
2
Landm arks and Semilan dmarks
Landmarks are discrete anatomical loci that can be recognized as the same point in all
specimens in the study. They are often termed “homologous points” because these points
can be matched up, one by one, as “the same point” in all individuals in the study. For
example, the mental foramen of the squirrel lower jaw ( Figure 2.1A ) is a discrete point
which corresponds to the mental foramen on any other mammalian jaw. “Homology” may
seem like the crucial term in the definition of a landmark, but the idea that these are dis-
crete points is equally crucial and these are notably sparse on some structures, including
the jaw. We could restrict the analysis to these few points ( Figure 2.1B .), but we would
miss a great deal of the morphology if we did. How much we would miss becomes more
obvious by removing the photograph from the picture and looking only at landmarks
( Figure 2.1C ). Landmarks are scarce, particularly in the regions of greatest interest, such as
where the muscles insert on the jaw. The complex curves that contain critical information
about the morphology motivates including additional points that can capture information
about curvature ( Fig. 2.1D ). These additional points are not discrete anatomical loci (much
less homologous); rather than being landmarks these points are “semilandmarks”.
There are some obvious differences between landmarks and semilandmarks. As noted
above, one difference is their discreteness, another is their homology. Related to these
is another: the position of semilandmarks along the curve is arbitrary
we could just as
easily sample the curve according to the scheme shown in Figure 2.2A or 2.2B . The ques-
tion every researcher must face is whether to include semilandmarks at all and, should
they decide “yes”, then the next question is how to sample them. How these questions
will be answered depends on the weight given to each of the several criteria for selecting
landmarks. That is the primary topic of the present chapter. Throughout much of our
discussion of these criteria, we will talk about selecting landmarks one by one, treating
each as an individual point, but landmarks (and semilandmarks) are not analyzed or inter-
preted one by one. The information about shape is contained in the entire constellation
of landmarks and semilandmarks, i.e. the configuration of points. That configuration is a
single datum. This may seem intuitively obvious to many readers, but readers experienced
in traditional morphometrics may find this view of a configuration as a single datum
counterintuitive because, in traditional morphometrics, each individual measurement was
often viewed as a trait in its own right. Even when analyzed multivariately, the individual
 
Search WWH ::




Custom Search