Morphometry offers new approaches for in vivo characterization of many neurologic and psychiatric pathologies. A survey of recent publications only hints at the attractiveness of magnetic resonance-based morphometry: published findings are heterogeneous, partly contradictory, and not always plausible in terms of known neuropathologic correlates. Hence, the sensitivity of the applied methods should be questioned. Three parameters affect the variance in morphometric findings: (1) knowledge about normal morphologic variability, (2) confounding physiologic parameters, and (3) methodologic misuse. Sound knowledge about the morphologic variability of the normal brain is vital for the assessment of volumetric findings. Large morphologic variability may also interfere with the precision of morphometric methods. The multitude of possible confounding physiologic parameters raises the necessity of precise subject control. Magnetic resonance scanning artefacts require rigid protocols, and application of the rather complex and sensitive methods demands profound insight into the techniques.