The anatomic accuracy of noninvasive in vivo measurement of tibial torsion was investigated through a comparison of goniometer measurements with those made on computed tomographic images. Seven normal subjects (2 women and 5 men; 14 legs) who ranged in age from 26 to 73 years were studied. The findings indicated that there was good agreement between measurements made by the two methods on the same limb. However, structural inconsistencies were found that cast doubt on the validity of certain anatomic reference points traditionally used in vivo studies of tibial torsion. In particular, use of the tibial tuberosity as a proximal reference may not give a true measurement of tibial or tibiofibular torsion.