The Variation menu contains commands for analyses of variation in samples.
The most widely used method is Principal Component Analysis (PCA). It is a useful tool to display variation within a sample and to characterize the main features of shape variation. If the data contain different subgroups, PCA can be used as an ordination method, but users should be aware that PCA is not optimized to find differences among groups (for that purpose, consider canonical variate analysis).
Matrix Correlation is an overall measure of the similarity of two covariance matrices. It is implemented with a test against the null hypothesis of no relationship between the covariance matrices, which uses the matrix permutation approach.
Contrast Covariance Matrices implements a test approach that focuses on differences between covariance matrices.
Procrustes ANOVA is a method for assessing the relative amounts of variation among individuals, of asymmetry (if both sides of the specimens have been measured) and of measurement error.