Previous theoretical research on the characteristics of map readability [see MSRS, 1988, No. 4, pp. 276-283] is tested using several groups of subjects (al-ready familiar with map use) in order to ascertain the most important elements facilitating map readability or detracting from it. The results are modeled through a series of regression equations in an effort to determine the relationships between specific characteristics (number of symbol size classes, frequency of symbols within classes, background complexity) and map readability (speed and accuracy of symbol identification).