This paper presents results from three experiments which compared gestural, keyboard, and mouse/keyboard interfaces to a spreadsheet program. This is the first quantitative comparison of these types of interfaces known to the author. The gestural interface employed gestures (hand-drawn marks such as carets or brackets) for commands, and handwriting as input techniques. In one configuration, the input/output hardware consisted of a transparent digitizing tablet mounted on top of an LCD which allowed the user to interact with the program by writing on the tablet with a stylus. The experiments found that participants were faster with the gestural interface than with the keyboard or mouse/keyboard interface. In addition, subjects tended to prefer the gestural interface over the keyboard interface. Inexperienced mouse users also tended to prefer the gestural interface over the mouse/keyboard interface, although experienced mouse users preferred the mouse. The main difficulties with the gestural interface had to do with poor display legibility and problems with the stylus. The benefits of the gestural interface are explained in terms of the fewer number of steps required to carry out an operation, the greater ease of remembering gestural commands, and the ability to focus on a single surface for input and output. © 1992 Taylor & Francis Group, LLC.