The eigenvalues of the covariance matrix of multivariate errors having anelliptical distribution are estimated by quadratic estimators. The availabledata have a p-way classification, have been aggregated, and may include(multivariate) outliers. This type of data may be encountered in several situations, including railroad or telephone demand studies or other network analyses. The main result of this paper is to provide explicit and straightforward-to-implement estimators of the.eigenvalues, not only when we simply assume that the distribution of the errors belongs to the class of elliptical distributions, but also in models where a particular elliptic distribution (such as the multivariate Student distribution, or the multivariate normal contaminated model) is specified as an alternative to the normal