Two studies aimed at defining radon prone areas in Toscana and Veneto, based on indoor measurements, are merged in this paper to extract the key points for a possible general methodology. In territorial units, the geometric mean (GM) and standard deviation (GSD) are evaluated from empirical data and the fraction of dwellings above the reference level (RL) is derived according to a log-normal model. In Toscana the sampling base is made of nearly 1,000 measurements (dwellings and kindergartens), apportioned according to the lithological classes (22) of the region. In Veneto 1,230 measurements in dwellings of the upper part of the region (estimates are given for the south), have been assigned to 6 x 5 km2 sequential units of the territory: the detection floor has been considered and results are referred to the regional mean dwellings condition with respect to the floor. Empirical GM and GSD values are corrected to increase statistical power based on two different methods. Maps of the percentages of dwellings with more than 200 Bq x m(-3) are presented. Areas with magmatic rocks in southern Toscana and northern parts of Vicenza and Belluno provinces in Veneto clearly emerge with more than 10% of dwellings above 200 Bq x m(-3). Efforts are required to establish proper confidence limits for estimates and to trace radon maps upon an administrative basis.