The 'pad' module performs discriminant analyses, permutation-based statistical significance of Mahalanobis ditances, reclassifications and validated ("leave one out") reclassifications, regression of discriminant factors on size (allometry). If you enter principal components (PC), the number of PC is adjusted to the smallest group of your data. ------------------------------------------------- INPUT FILES ------------------------------------------------- Traditional MM --------------- Size in <- Log-measurements between landmarks Size out <- Log-measurements between landmarks --------------- Geometric MM --------------- Total shape <- Shape variables (..._PW,..._FOGnef,..._HAMPnef) Partial shape <- Principal components of shape variables (..._RW, ..._FOGnefcp) --------------- Others --------------- Raw data <- Any file respecting the CLIC data format Meristic traits <- Any file respecting the CLIC data format ------------------------------------------------- OUTPUT FILES ------------------------------------------------- ending by: ..._PW_MAHAphylip.txt ------------------------------------------------- The CLIC data internal format is as follows: - The first row is a free comment, next rows are the data. - The first row is mandatory. - Data (next rows) are organized as a table without headers - Empty rows are allowed, except the first one (see above). - Columns are separated by a tabulation. Exceptions to this CLIC data format are 1. The 'coord_[date].txt' file, which is the natural 'coo' output file: it has an internal format close to the TPS format of Rohlf. 2. The CLIC data base format ('..._OTL_DB.txt'): there are headers! 3. The CLIC distance matrix format ('..._MAHAphylip.txt', etc.): except for the first row (the comment), its format depends on the external software to which it is destinated.