The sources of our knowledge are twofold. In a few cases archaeological excavation has laid bare the paving of Roman streets or the foundation of Roman house-blocks. More often mediaeval and modern streets seem to follow ancient lines and the ancient town-plan, or a part of it, survives in use to-day. Such survivals are especially common in the north of Italy. It is not, indeed, possible to gather a full list of them. He who would do that needs a longer series of good town-maps and good local histories than exist at present; he needs, too, a wider knowledge of mediaeval Italian history and a closer personal acquaintance with modern Italian towns, than a classical scholar can attempt. But much can be learnt even from our limited material.[68] [68] See the seventeenth century Atlases of Blaeu, Janssons, and others, the modern maps prepared by Grassellini and others about 1840-50 (some on the scale 1:4,000), and in particular the _Atlante geografico_ of Attilio Zuccagni-Orlandini (Firenze, 1844), and the recent town-maps of various Italian cities (mostly about 1:10,500). Different maps of the same town sometimes differ much in their detail. The Italian Government maps of the largest scale (1:25,000) are small for our present purpose and have been issued mainly for northern Italy. The evidence of the streets needs, however, to be checked in every case. It would be rash to assume a Roman origin for an Italian town simply because its streets are old and their plan rectangular. There are many rectangular towns of mediaeval or modern origin. Such is Terra Nova, near the ancient Gela in Sicily, built by Frederick Stupor Mundi early in the thirteenth century. Such, too, Livorno, built by the Medici in the sixteenth century. Such, too, the many little military colonies of the Italian Republics, dotted over parts of northern and middle Italy. Often it is easy to prove that, despite their chess-board plans, these towns do not stand on Roman sites. Often the inquiry leads into regions remote from the study of ancient history.