Experiments on fifteen simulated keyboards with different key sizes and inter-key spacings are reported. It was found that the movement time on these keyboards was well described by the model of Drury (Drury and Hoffmann 1992). Minimum movement times occurred when the inter-key spacing was approximately equal to the finger pad size. There was found to be an effect of the number of keys moved, which was not predicted by the model. This effect remained in a further experiment in which the target keys were marked.
PDF full text from Publisher