It is unclear how to interpret ulnar variance (UV) as determined by magnetic resonance imaging (MRI). Using a radiology database, we retrospectively assessed UV on MRI and compared it with UV on radiographs. MR images of 163 wrists (158 patients) were reviewed. Mean (SD) UV was -0.16 (2.43) mm on radiographs, -0.62 (2.41) mm on T1-weighted (bone-to-bone) MRI, and -0.50 (2.38) mm on gradient- echo or short tau inversion recovery (cartilage-to-cartilage) MRI. Compared with MRI, radiographs significantly underestimated the magnitude of negative UV. There was no difference in UV between different hand positions in MRI. Further research is needed to validate the measurement of UV on MRI and to determine its clinical utility.