Recent evidence for the use of visual cues in the programming of the precision grip has been given by Gordon et al. (1991). Visually invoked size-related information influenced the physical forces used to produce a lift, even when it was not consistent with other sensory information. In the present study, blind-folded subjects were required to feel the size of an object by haptic exploration prior to lifting it. Two boxes of equal weight and unequal size were used for the lift objects and were attached to an instrumented (grip) handle. Grip force and load force, their rates, and the vertical movement of the object were measured. Most subjects reported that the small box was heavier, which is consistent with size-weight illusion predictions. However, peak grip force, grip force rate, peak load force, and load force rate were greater for the large box when the boxes were randomly presented, but not when the same boxes were lifted consecutively. If subjects did not feel the box prior to a lift, these parameters were scaled in between those normally employed for the large and small box. Most subjects apparently programmed the parallel increase of the grip and load force during the loading phase as one force rate pulse. This represented a "target strategy" in which an internal neural representation of the objects weight determined the actual target parameter (i.e. just enough force required to overcome gravity). The other subjects exhibited a slower stepwise increase in grip and load force rate. The subjects choosing this "probing strategy" did not scale the force parameters differently for the two boxes.(ABSTRACT TRUNCATED AT 250 WORDS)