Dillemann (1979, 63) notes that “names at the end of a chapter may be additions having no bearing on the road network”. Apart from the fact that the main body of the text for Britain is not based on the road system anyway, if we search the text for additions, it will be seen that they occur in one place only. Not at the end of Britain south of Hadrian’s Wall, taken by Rivet and Smith (1979, 193) as marking the end of the Cosmographer’s main source, nor at the end of the first ‘section’ of twenty-four names, so far as can be ascertained; the only real addition is the list of diuersa loca, which follows the list of placenames proper.
With river names, the situation is slightly different. Having travelled clockwise around Britain, omitting names from north-western Scotland apparently because he mistook them for island names (hence the river names at 10914 and 10920 in the island list), the Cosmographer seems to have reached Kent and noticed that he had left out several names from north-eastern England, which he then inserts at that point (10837-38); right at the end of the listing he inserts at least one extra name (*Uxela, the River Parrett, 10842). The island lists are too obscure to be certain of any method that the writer may have used, although I have suggested that he was working in discrete groupings rather than jumping randomly around the west coast, as the identifications proposed by Rivet and Smith imply.
Although this does not form absolute proof of the unity of the Cosmographer’s source, it is unlikely that an individual who was clearly a muddled thinker could have grasped the contents of three or more maps simultaneously and returned to them to list names he had omitted. Far from it: he repeats names in situations which cannot indicate the use of multiple sources (e.g. *Leuiodunum at both 1089 and 10810), and does not recognise names he ought to know (such as <Staurinis> at 6633 for Taurinis, Turin in his native northern Italy, quoted by Rivet & Smith 1979, 187). If one source will explain the nature of the British section, as I have shown, let us accept it rather than multiply sources ad infinitum.