378 TEXTUAL ERRORS IN THE July the aggregate numbers of the stages. On the other hand, there is reason to suspect that accidental errors and intentional altera- tions crept into the text during the fourth century. For large portions of the Itinerarium we have two manuscripts, the Escoria- lensis P and the Parisinus D, which differ sufficiently from the rest of our manuscripts to lead us to believe that, at least, three separate recensions of the Itinerarium were current in the fourth century. Dr. W. Kubitschek ^ has tried to prove that the Itinerarium was exclusively derived from a road-map of the world, and that the same road-map was also the ultimate source, not only of our existing road-map known as the Tabula Peutingeriana, but also of the curious compilation of the cosmographer of Ravenna. Dr. Kubitschek's arguments, so far as they imply a generous use of a road-map by the compiler of the Itinerarium and a connexion between the map so used and both the Tabula and the work of the cosmographer of Ravenna, are cogent : but I do not think he has proved that the compiler of the Itinerarium used no source other than the road-map. We have seen that, of the three pre-Carolingian recensions of the text of the Itinerarium which are known to us, one only is available for Britain. The Tabula Peutingeriana is equivalent to a fourth recension, but it contains no more of Britain than the eastern counties, Kent, and Devon- shire. In these parts of Britain, therefore, and especially on pp. 479-80, in the ninth Iter, our text has twice as broad a basis as elsewhere. Before the Tabula can be used, it is necessary to remove some imperfections which are due to the fact that the names near the margin of the map are without their initial letters : for Ridunio, we must read Muriduno ; for Madus, Noviomagus ; for Rotibis, Durotibis = Durobrivis ; for Baromaci, Caesaromagi ; and for ad Taum, Venta. The last of these corrections may, at first sight, appear less obvious than the others ; but ad Taum stands above one of those vignettes which are used on the Tabula to indicate either a capital of a canton, or a legionary camp.^ It is not credible that the Icenians had a capital ad Taum in addition to their capital Venta ; and, consequently, we are bound to assume that all but two letters of ad Taum are conjectural, and that the first three letters of Venta ^ were lost in the same way as the first three ' See Eiru romiache Straasenkarte (Jahreshefte des Oesterreichischen Archdologischen Institutes, v. 1902, 31).
- See Desjardins, Oiographie de la OatUe, iv. 107.
• An eleventh-century writer, William of Poitiers, who was fond of airing his classical knowledge, mentions Norwich under the name of Gventa. He may have obtained this information from the map from which our Tabula was copied. If so, ad Taum is a thirteenth-century corruption. (See Migne, Patrologia Lalina, cxliz. 1263 ; and Freeman, Norman Conquest, iv (2nd ed. 1876), p. 67, n. 2.)
