Evidence has recently been given by Gordon et al. (1991a, b) for the use of visually and haptically acquired information in the programming of lifts with the precision grip. The size-related information influences the development of manipulative forces prior to the lift-off, and the force output for larger objects is adjusted for a heavier weight even if the weight of the objects is kept the same. However, the size influences on the force output were small compared to the relative effects of the expected weight in previous trials (Johansson and Westling 1988). In the present study, both the size and weight of objects were changed between consecutive lifts to more fully determine the strength of visual size cues. During most trials, the size and weight covaried (i.e. the weight was proportional to the volume). However, in some trials, only the size was switched while the weight was kept the same to create a mismatch between the size and weight. The forces were still appropriately scaled towards an expected weight proportional to the volume of the object. It was concluded that visual size cues are highly purposeful. The effects were much larger than previously reported and were similar in magnitude to the effects based upon the expected weight. Thus, the small effects reported in the previous experiments may have been a result of conflicting "size-weight" information.