We investigated the relative effectiveness of tangential versus normal displacements of skin for producing tactile sensations. Subjects adjusted the magnitude of slow tangential and oblique displacements of a flat-ended, cylindrical, 1 mm diameter probe glued to the skin in order to match the perceived intensity of a reference displacement that indented the skin normal to the surface. At both the forearm and fingerpad, subjects chose tangential displacements only 0.3 to 0.6 times as large as the reference normal displacement, indicating a significantly higher sensitivity to tangential displacement. Based on measurements of the mechanical impedance of the skin to normal and tangential displacements, these results were also expressed in terms of forces. At the forearm, subjects were more sensitive to tangential forces than normal force. However, at the fingerpad, sensitivity to tangential forces was lower than sensitivity to normal force, due to the approximately five-fold greater stiffness of the fingerpad to tangential traction. These results provide guidance for development of tactile displays: (1) When an actuator is limited primarily in terms of peak displacement (e.g., the maximum strain of a ceramic peizoelectric actuator) then tangential stimulation is a superior choice for both body sites we tested. (2) When an actuator is limited primarily in terms of peak force (e.g., the stall torque of a DC micromotor) tangential stimulation is the superior choice for the hairy skin, but normal stimulation is the better choice on the fingerpad.