We analyzed the official yearly mortality data due to malignant pleural mesothelioma (MPM) from 1995 to 2002 in all 107 Italian provinces, and we related them to provincial resident population in the same years. The resulting ratios are displayed as geographically referenced maps, which enhance the visual impact on how that fatal disease distributes over the country. The following results are statistically meaningful: (1) internal migration changed the Italian population distribution, with major increases in regions Lombardy, Trentino-Alto Adige, Veneto, and Emilia-Romagna, and moderate decreases in Piemonte, Liguria, and Friuli-Venezia Giulia.(2) MPM mortality is increasing overall, with five provinces largely exceeding the average Italian rate: Alessandria (Piemonte), Genoa and La Spezia (Liguria), and Gorizia and Trieste (Friuli-Venezia Giulia). (3) There is a substantial mortality rate increase in the same provinces which shows decreasing population. Results are discussed with reference to well-known cases of asbestos-polluted sites such as Balangero and Casale Monferrato (Piemonte), Genoa and La Spezia (Liguria), and Monfalcone and Trieste (Friuli-Venezia Giulia). General unquestionable conclusions are (a) occupational exposure, especially working conditions, increases MPM mortality; (b) non-occupational exposure data are scattered, as are male versus female data, and cannot be dealt with statistically on a nationwide scale; (c) natural environmental exposure may contribute to MPM mortality, but its effect is still statistically minor. The MPM mortality apex is currently reached at Trieste, a province free of asbestos minerals occurring in their natural setting, but where MPM affects workers in the marine industry strongly who operated in contact with insulating outfits. Information on content and type of amphibole present in the asbestos materials used for industrial applications is scattered and generally scarce, thus making the evaluation of the amphibole versus chrysotile relative impact on MPM mortality rates statistically unsound.