ABSTRACT
Introduction ............................................................................................................ 144 Images, Three-Dimensional Scans, Morphometric Ordination and Neural Nets .... 147
Digital Images............................................................................................. 147 Three-Dimensional Scans and Virtual Specimens .................................... 151 Morphometric Ordination ........................................................................... 152 Artificial Neural Nets (ANNs) .................................................................... 158
A Comparative Analysis ........................................................................................ 161 Materials and Methods .......................................................................................... 161
Materials ..................................................................................................... 161 Data Collection Methods ............................................................................ 162
Digital Photography ......................................................................... 162 Three-Dimensional Laser Scanning ................................................ 164
Data Analysis Methods ............................................................................... 166 Three-Dimensional Relative Warps Analysis .................................. 166 Eigensurface Analysis ...................................................................... 173 PSOM ANN Analysis ...................................................................... 178
Results .................................................................................................................... 179 Three-Dimensional Relative Warps Analysis ............................................ 179 Three-Dimensional Eigenshape Analysis................................................... 181 Eigensurface Analysis ................................................................................ 183 PSOM-ANN Analysis ................................................................................ 189
Discussion .............................................................................................................. 196 Summary ................................................................................................................ 201
There are many definitions of morphometrics, but perhaps the broadest is ‘the associations, causes, and effects of form’ (Bookstein, 1991). Over the last 50 years morphometric methods have been instrumental in advancing the study of how organismal shape covaries with size (allometry), time (evolution), environment (ecophenotypy), geography (morphological biogeography), ecology (ecomorphology), development (heterochrony), and a host of related factors (see Sokal and Sneath, 1963; Blackith and Reyment, 1971; Reyment, 1971; Sneath and Sokal, 1973; Pimentel, 1979; Reyment, 1980; Reyment et al., 1984; Reyment, 1991). More recently, the school of geometric morphometrics1 has added further rigor to this research programme by synthesizing deformational (e.g. Siegel and Benson, 1982) and multivariate (e.g. Blackith and Reyment, 1971; Bookstein, 1978) approaches to shape comparison and placing both on a firmer mathematical footing (see Kendall, 1977, 1984; Mardia and Dryden, 1989; Rohlf and Bookstein, 1990; Bookstein, 1991). Books, technical articles, conference abstracts and student theses that employ morphometric approaches now abound (Adams et al., 2004). How odd then that morphometrics in general and geometric morphometrics in particular have largely neglected the most fundamental biological correlate of form-the correlation on which virtually all other morphometric studies critically depend and that provides the most obvious link between morphometrics and mainstream biology: the correlation between form and taxonomy.
