Theoretical studies on metal cations in water have lead to a controversy regarding the distance between the cation and the oxygen nuclei of the first solvation shell. Apparently this is due to the differences in the description of solvent effects which can lead to opposite conclusions: a shortening or a lengthening of this distance in the bulk solution with respect to that found in the hydrated cation in a vacuum. This discrepancy has been attributed to the difference between discrete and continuum solvation models. Likewise, within the continuum model, two widely used methods, those of the multipole expansion (MPE), from the Nancy group (using a spherical cavity), and the polarized continuum model (PCM) from the Pisa group (using a molecularly shaped cavity), have given different results. This work reconsiders the solvent effects on the geometry of a set of hydrated cations ([M(H2O)(m)](n+), where M = Be2+, Mg2+, Ca2+, Zn2+, and Al3+) by comparing results derived from a previous work [J. Phys. Chem. 1991, 95, 8928] using the MPE approach with new PCM calculations. It is shown that both methods lead to the same results when the same (spherical) cavity is used, but PCM results, with molecularly shaped cavities, are in better agreement with those obtained with discrete solvent models. A continuous change in the atomic parameters defining the cavity allows the transition from one cavity shape to another and to observe the reversal in the prediction of the change in the M-O distances. The present study shows the substantial equivalence of the two methods (never accurately checked in precedence) and reveals that the M-O distance in the first hydration shell is an appropriate parameter to monitor finer aspects of the solute-solvent interactions, related to the discreteness of the solvent interactions in the second shell. These effects can be reproduced at a very good extent by continuum models with molecularly shaped cavities.