Unexpected pulling and pushing loads exerted by an object held with a precision grip evoke automatic and graded increases in the grip force (normal to the grip surfaces) that prevent escape of the object; unloading elicits a decrease in grip force. Anesthesia of the digital nerves has shown that these grip reactions depend on sensory signals from the digits. In the present study we assessed the capacity of tactile afferents from the digits to trigger and scale the evoked grip responses. Using tungsten microelectrodes inserted percutaneously into the median nerve of awake human subjects, unitary recordings were made from ten FA I and 13 FA II rapidly adapting afferents, and 12 SA I and 18 SA II slowly adapting afferents. While the subject held a manipulandum between a finger and the thumb, tangential load forces were applied to the receptor-bearing digit (index, middle, or ring finger or thumb) as trapezoidal load-force profiles with a plateau amplitude of 0.5-2.0 N and rates of loading and unloading at 2-8 N/s, or as "step-loads" of 0.5 N delivered at 32 N/s. Such load trials were delivered in both the distal (pulling) and proximal (pushing) direction. FA I afferents responded consistently to the load forces, being recruited during the loading and unloading phases. During the loading ramp the ensemble discharge of the FA I afferents reflected the first time-derivative of the load force (i.e., the load-force rate). These afferents were relatively insensitive to the subject's grip force responses. However, high static finger forces appeared to suppress excitation of these afferents during the unloading phase. The FA II afferents were largely insensitive to the load trials: only with the step-loads did some afferents respond. Both classes of SA afferents were sensitive to load force and grip force, and discharge rates were graded by the rate of loading. The firing of the SA I afferents appeared to be relatively more influenced by the subject's grip-force response than the discharge of the SA II afferents, which were more influenced by the load-force stimulus. The direction in which the tangential load force was applied to the skin influenced the firing of most afferents and in particular the SA II afferents. Individual afferents within each class (except for the FA IIs) responded to the loading ramp before the onset of the subject's grip response and may thus be responsible for initiating the automatic increase in grip force. However, nearly half of the FA I afferents recruited by the load trials responded to the loading phase early enough to trigger the subject's grip-force response, whereas only ca. one-fifth of the SA Is and SA IIs did so. These observations, together with the high density of FA I receptors in the digits, might place the FA I afferents in a unique position to convey the information required to initiate and scale the reactive grip-force responses to the imposed load forces.