-------- Original Message -------- Subject: Question: depict shape changes associated with CV's Date: Fri, 10 Oct 2008 11:50:35 -0700 (PDT) From: Rebeca <[EMAIL PROTECTED]> To: <[email protected]> Dear Morphometricians, I am trying to assess shape variation in a fish species coming for four different geographical areas using landmark-based GM. To do this, I used the Partial warp scores in a MANOVA and CVA to see group differences. I would like to depict the shape changes associated with canonical variates using deformation grids with TPSRegr software, as described in Rohlf et al. 1996 (Morphometric Analysis of Old World talpidae Using Partial-Warp Scores) but I am not sure how to do it. I understood that I should regress the partial warp scores onto the canonical scores of each CV axis. For example, I calculated the CV1 scores (unstandardized canonical scores) using STATISTICA and then I prepared a NTS file to use as the independent variable in the TPSRegr software. However, when I try to do the regression, I get an error (“Bad value for independent variable on specimen 1”…). Many of the scores are negative values, could this be the problem? Am I regressing the wrong variables? I will appreciate a lot any help with this issue. Thanks a lot, Rebeca Rodriguez Rebeca P. Rodriguez Mendoza PhD student e-mail: [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]> Marine Research Institute – Fisheries Dept. C/ Eduardo Cabello, 6 36 208 Vigo, Spain -- Replies will be sent to the list. For more information visit http://www.morphometrics.org
