Questions of value discrimination, value estimation, and regionalization were asked of a sample of map users from a set of computer-generated choropleth maps. The objectives of the study were to determine whether map-user responses differed between classed and unclassed maps, between crossed-line and solid-filled symbology, and with different numbers of data units on the maps. Classed choropleth maps performed significantly better than unclassed in value estimation, whether the symbolization was crossed-line or solid-filled. However, there was no significant difference between them in regionalization tasks. For symbolization, crossed-line generally performed better than solid-filled, significantly better in value-discrimination and value-estimation tasks, but not in regionalization. Generally, response patterns did not change with increasing numbers of data units on maps. However, there was a significant difference for such changes with unclassed maps using solid-filled symbols for value-discrimination an...