Procrustes analysis involves finding the optimal superposition of two or more ''forms'' via rotations, translations, and scalings. Procrustes problems arise in a wide range of scientific disciplines, especially when the geometrical shapes of objects are compared, contrasted, and analyzed. Classically, the optimal transformations are found by minimizing the sum of the squared distances between corresponding points in the forms. Despite its widespread use, the ordinary unweighted least-squares (LS) criterion can give erroneous solutions when the errors have heterogeneous variances (heteroscedasticity) or the errors are correlated, both common occurrences with real data. In contrast, maximum likelihood (ML) estimation can provide accurate and consistent statistical estimates in the presence of both heteroscedasticity and correlation. Here we provide a complete solution to the nonisotropic ML Procrustes problem assuming a matrix Gaussian distribution with factored covariances. Our analysis generalizes, simplifies, and extends results from previous discussions of the ML Procrustes problem. An iterative algorithm is presented for the simultaneous, numerical determination of the ML solutions.heteroscedasticity ͉ morphometrics ͉ Procrustes analysis ͉ superpositions ͉ least-squares T he goal of Procrustes analysis is to superpose non-identical shapes in an optimal manner via scalings, translations, and rotations (1-3). Classical Procrustes analysis uses the unweighted least-squares (LS) criterion for finding the optimal transformations. LS implicitly assumes that the landmarks describing the forms' shapes are uncorrelated and that they have identical variances (i.e., that they are homoscedastic). However, in many practical applications these assumptions are known to be violated. For instance, when superpositioning macromolecular proteins, individual atoms (the landmarks) are connected via covalent chemical bonds, and thus the variance of a given atom can be correlated with the variance of atoms to which it is connected. Furthermore, certain atoms may have larger mobilities than others, or they may have spatial positions with relatively greater uncertainty due to experimental error. Similarly, when comparing the skulls from different members of a species, some homologous features may be highly variable relative to others. Hence, different landmarks can have widely different variances. Under these conditions, ordinary LS can give misleading results, even with large samples of data (4).The method of maximum likelihood (ML) is a common alternative to LS and is widely considered to be fundamental for statistical modeling and parameter estimation (5). Given the proper model, ML can provide accurate and robust estimates of parameters in the presence of both heteroscedasticity and correlation. Incorporating heterogeneous variances and non-zero correlations into the ML Procrustes problem involves weighting by two covariance matrices (a landmark and a dimensional covariance matrix). However, estimation of these covariance matrices...