380 TEXTUAL ERRORS IN THE July xii instead of xu after Londinio, and xu instead of xii after Caesaro- mago. If we compare the numbers given by the Itinerarium and by the Tabula, we shall perceive that the same confusion of u and ii has taken place in the stage from Sitomagus to Combretonium. We must read xii instead of xu in the Tabula. In other respects the numbers of the Tabula for this and for the preceding stage are clearly preferable to those of the Itinerarium, which has an x wrongly prefixed to each number. The fact that the heading number of the ninth Iter agrees with the inflated numbers is not to be taken as a confirmation of the latter, since our text affords abundant evidence that intentional alterations were made for the purpose of bringing the heading numbers into conformity with the aggre- gate numbers of the stages. Of our best manuscript, the Vindo- bonensis L, the editors of the Itinerarium, Parthey and Pinder, remark : * in hoc libro praestantissimo numeri collecti, qui singulis itineribus anteponuntur, fere omnes erasi et secunda manu correcti sunt.' * Paradoxical as it may seem, the heading numbers, where they agree with the aggregates of the stages, are of poor authority ; where they differ, they often have considerable critical value. The numbers of the first two stages being corrected to xxii and xii, Sitomagus will be 22 miles from Venta, that is, near the place where the road crosses the River Dove at Stoke Ash. The aggregate distance from Venta to Camolodunum will be 55 miles according to the ninth Iter and 53 miles according to the fifth. It appears, however, that the compiler, or, more probably, a redactor, was aware that two miles had been lost somewhere in the fifth Iter, for he has added two miles to the last stage of the Iter, thus making the distance from Verterae to Lugu- vallum 42 miles instead of 40. The correct distance (40 miles) is given at p. 467, in the second Iter {Luguvallo, xiiii Voreda, xiii Brovonacis, xiii Verteris). A similar alteration (in this case a sub- traction of two miles) has occurred on p. 477, at the end of the sixth Iter, where the last stage {Crococalana to Lindum) is given as 12 miles instead of 14, which is the number given on p. 478, in the eighth Iter. The reason for this subtraction of two lies in the fact that, in the sixth Iter, the distance from Bannaventa to Ven^nae is wrongly given as 20 miles instead of 18 miles, which is the number given on p. 479, in the eighth Iter, or 17 miles, which is the number given on p. 470, in the second Iter. It seems clear that the compiler, or a redactor, sometimes altered his numbers, preferably the number of the last stage, in order to make his aggre- gate agree with his heading number. It is certain that the com- piler of the Itinerarium often obtained his heading numbers from
- Itinerarium Ant. Aug., edd. Parthey and Pinder (Berlin. 1858), p. xviii.
