I agree with the idea of Dr. Warheit. And also apologize for my ingorance and possibly naive statements, nevertheless I would like to add some coments on the same theme.
I have been thinking on how to remove allometric relationships after applying warp transformations, and probably working with the residuals from a regression on centroid size from parameters such as relative warps will represent more accurately phenomena related to shape. Should these residuals will represent more accurately evolutionary relationships than the original data? At least the effects of size (allometry and isometry) are reduced. I understand that relative warps can be regressed individually on centroid size since they are independent. Are these new set of variables (residuals) more informative about evolutionary change than ortogonal projections of the same data on the first principal component obtained from continuos characters (e.g. longitudinal measurements in skull and body bones)? I suppose this is correct, since the rest of components contain useful information (but they can also be "contaminated" by "size" variation) and they are usually not considered together with the projected data; moreover, the variables originated from landmarck data usually contain more information than traditional continuous characters. Thank you for all your patience and advice. Pablo Pablo Jarr�n Center for Ecology and Conservation Biology Biology Department Boston University [EMAIL PROTECTED] == Replies will be sent to list. For more information see http://life.bio.sunysb.edu/morph/morphmet.html.
