Statistical Methods for Comparison of Forms - An Invariant Approach to Statistical Analysis of Shapes - page 188

Figure 4.10. Differences in mandibular morphology between Ts65Dn aneuploid mice
and their normal littermates. Thin lines indicate those linear distances that are small-
er in the aneuploid individuals, while the bold lines indicate those linear distances
that are larger in the aneuploid individuals. Interpretation of these results are that
the morphology of the coronoid process (landmark 1, indicated by arrow) has changed
such that the tip has moved more anteriorly and inferiorly.

morphology of the coronoid process has occurred that results in reduc-
tion of the overall process and anterior migration of its apex. Such a
local change in morphology can account for the specific pattern of dif-
ferences in linear distances on the mouse mandible.

The delete-one landmark approach

Another way to determine the influence of a particular landmark on
the overall form difference is to re-run the EDMA-I or EDMA-II anal-
ysis after deleting one or more landmarks. Since we have found
landmark 1 to be at both the minimum and maximum extremes of the
form difference matrix calculated using all eleven landmarks, we run
the EDMA-I analysis again, after deleting landmark 1.
Inspection of the form difference matrix indicates that the number
of linear distances that are less than 1 has decreased. Below, we write
the form difference matrix as a vector and observe the minimum and
maximum values.

Form Difference matrix with landmark 1 deleted written as a
sorted vector:
LND 6

LND 5 0.98568

LND 10 LND 5 0.99466

LND 10 LND 9 1.00052

LND 11 LND 5 1.00662