Objective To investigate ankle torque and steadiness in the intact leg of transtibial and transfemoral unilateral amputees. Design Comparative study. Setting Medical rehabilitation centers. Participants Fifteen persons with a unilateral transfemoral amputation, 8 persons with a transtibial amputation, and 14 able-bodied male participants volunteered to participate in this study (N=37). Interventions Not applicable. Main Outcomes Measures Maximal isometric torque performed during ankle plantarflexion and dorsiflexion in the intact limb of amputees and in the dominant limb of able-bodied persons. The coefficient of variation (CV) of the plantarflexion torque was calculated over 5 seconds during a submaximal isometric contraction (15%) in order to assess torque steadiness. Furthermore, electromyographic activity (the root mean square amplitude) of the gastrocnemius medialis and tibialis anterior muscles was analyzed. Results Plantarflexion maximal torque was significantly higher for the able-bodied group (115±39 Nm) than for the group with a transfemoral amputation (77±34 Nm) (P<.01), and did not differ between able-bodied group and the group with a transtibial amputation (97±26 Nm) (P=.25). Furthermore, the transfemoral amputee group was 29% less steady than the able-bodied group (P=.01). However, there were no significant differences in torque steadiness between the able-bodied group and transtibial amputee group (P=.26) or between transtibial and transfemoral amputee groups (P=.27). The amputation had no significant effect between groups on dorsiflexion maximal torque (P=.10), gastrocnemius medialis electromyography (EMG) (P=.85), tibialis anterior coactivation (P=.95), and coactivation ratio (P=.75). Conclusion The present study suggests that as the level of amputation progresses from below the knee to above the knee, the effect on the intact ankle is progressively more negative.