Map and geospatial information is an essential tool in many fields of human activities. Although cartographers were extensively involved in cognitive map design research during the second part of the 20th century, the map use process as well as the significance of individual map design elements have not been studied in sufficient depth. The article discusses the results of a controlled experiment. The exploratory study aimed to simulate the execution of some problem-solving tasks based on the information derived from thematic maps with differently designed legends, viz.: list-legend, grouping-legend and natural-legend. On the basis of think-aloud protocols, it has been confirmed that some differences in map reading depend on the legend used.