The great diversity of prosthetic mechanisms available nowadays leads to the question of which type of artificial foot would be the most advisable for a particular person. To answer correctly, it is necessary to establish, in an objective way, the performance of each type of prosthetic mechanism. This knowledge is obtained by means of the study of the subject-prosthesis interaction, both in static and dynamic conditions. This paper, based on the analysis of 8 trans-tibial (TT) amputees, presents a quantitative method for the study of human gait which allows the determination of the influence of four different prosthetic ankle-foot mechanisms (SACH, Single-axis, Greissinger and Dynamic) on gait. To do this, 1341 gait trials at different cadences were analysed (383 with normal subjects and 958 with amputees, using the four prosthetic feet under study). From all the variables available for study only those which offered interpretable clinical information were chosen for analysis. A total of 18 variables (kinetic, kinematic and time-related) were selected. A covariance analysis (ANOVA) of these variables was made, which showed that the factors influencing TT amputee gait were, in order of importance, cadence and leg studied (sound or prosthetic), inter-individual variability and, finally, the prosthetic mechanism used. When looking at the performance during gait of the 4 prosthetic mechanisms studied it can be observed that there are similarities in the kinetic study between SACH and Dynamic feet on one hand and Single-axis and Greissinger feet on the other. These results seem to support the classification criteria of articulated and non-articulated prosthetic mechanisms.