Rivet and Smith (1979, 212) accept that this is basically a list of places north of the Antonine Wall, although their statement that &ldsquo;here no particular order can be expected, since there was no road-system” is somewhat overstating the case. There appear to have been two roads leading into northern Scotland: the well-known route through Ardoch, Strageath, Bertha, Cargill and Cardean (Margary 1973, 491), and a lesser-known road through Drumquhassle and Malling, possibly connecting with Bochastle and Dalginross. In addition, the forts along this western line must have been connected by road with those to the east.
If we accept that the Cosmographer and Ptolemy share a common source of information for the far north of Scotland, we are forced to accept that there were only three occasions in the first century when names of places could have been collected in this remote part of Britain: during the circumnavigation of Scotland by Demetrius in AD 81-3 mentioned by Plutarch (Jones & Keillar 1996, 47), during that mentioned by Tacitus (Agricola 38) as belonging to the seventh season of Agricola’s campaigns, and during the army’s advance along the east coast during that campaign (Agricola 29). Just as Ptolemy (or his source) appears to have made use of marching-orders or some similar source to give structure to the listing of places north of the Tyne-Solway line, so the Cosmographer’s map source probably also ordered its names according to a topographic sequence reflecting the order of collection (or coining) of names. We have already seen that apart from the second-century lists of forts on Hadrian’s Wall and the Antonine Wall, there is little in the Cosmographer’s text to suggest a source for the north of Britain later than the first century CE.
Using this hypothesis to structure the text, it can be divided into three sections, the first and third listed in anti-clockwise order, the second in clockwise order. The first section ends with an apparent mention of Ireland, presumably depicted at the south-western (or, rather, anti-clockwise) extremity of the map north of the Antonine Wall. The second section begins with *Pinnata Castra, the most distant place mentioned by Ptolemy. The third appears to begin with *Dexa, identified above with Camelon. The structure of this section therefore seems to be relatively straightforward: a listing of places not linked by road, but encountered during the Agricolan circumnavigation; a section listing places again not linked by road and occupied only temporarily during Agricola’s final season of campaigns; and finally, a section beginning in an area in which permanent forts, linked by roads, were established. It is possible that the Cosmographer’s map used different symbols for each class of place, or distinguished them in some other visible way, which would explain the order he imposed onto the material.