2004 Denver Annual Meeting (November 7–10, 2004)

Paper No. 7
Presentation Time: 1:30 PM-5:30 PM

REMOVING DEFORMATION FROM FOSSILS: SIMULATION TESTS WITH GEOMETRIC MORPHOMETRIC DATA


ANGIELCZYK, Kenneth D., Dept. of Invertebrate Zoology and Geology, California Academy of Sciences, 875 Howard St, San Francisco, CA 94103-3009, SHEETS, H. David, Dept. of Physics, Canisius College, 2001 Main St, Buffalo, NY 14208 and ROOPNARINE, Peter D., Dept. of Invertebrate Zoology and Geology, California Academy of Sciences, 875 Howard St, San Francisco, CA 94103, kangielczyk@calacademy.org

Many fossils undergo tectonic deformation, and the inclusion of such deformed specimens in morphometric analyses can result in misleading interpretations of morphologic data. Several techniques for removing the effects of deformation from fossils have been proposed, but few have been applied to landmark-based shape data in a geometric morphometric framework. Furthermore, none have been tested systematically with simulation studies. Here we report the results of simulation tests of four potential methods for removing the effects of deformation from landmark data.

The starting point for our simulations is a data set of 19 landmarks on 409 plastra of the extant turtle Emys marmorata, which shows a significant pattern of ontogenetic shape change. Landmark configurations were subjected to known amounts of uniform shear, the effects of which we attempted to remove using the four methods. The ontogenetic trajectories of the original and ‘un-deformed’ data sets were then compared to assess the success of each method. We considered a method successful if it could remove the effects of deformation without significantly altering the true ontogenetic signal.

The four methods tested were 1) reflection and simple averaging of the coordinates of a specimen’s bilaterally symmetric landmarks; 2) a generalized resistant fit affine superimposition of specimens; 3) a PCA-based modeling of the effects of deformation; and 4) a modified version of Motani’s method that models deformation using sets of bilaterally symmetric measurements and known symmetry relationships. Preliminary tests of methods 1 and 3 suggest they cannot remove the effects of deformation without altering the ontogenetic signal. This result is especially surprising for method 3 because it qualitatively seems capable of detecting a deformation signal. Developing a method to remove the effects of deformation is important because it broadens the range of data and questions accessible to geometric morphometrics.