8,5:3 Magnesia

Miller correctly argues that Trallis has been scripted a line too high, though he seems to miss a corroborating argument: Trallis is not inscribed in the crook of the chicane drawn under it, a factor which increases the suspicion that a copying error has occurred. The number XXVIII would plausibly pertain to the Trallis-Magnesia connection. What is obviously missing is a set of stations and distances covering the connection from Hierapolis to Tralles, which leads me to the conclusion that Tralles must have been drawn as a terminus.

 This clumsy situation seems to have arisen because two itineraries heading up the Maeander valley from Magnesia were employed by the chart-maker, who failed to recognize them as a single route and made the mistake of separating them into a wide V shape. The fact that his Cayster (Castvr) valley contains not only Hypaipa (ypepa), Anokome (angome) and Ephesus, but also Magnesia, instead of keeping Magnesia in the Maeander (Mindo) valley where it belongs, confirms a degree of confusion.

Miller's radical revirement (725-726 and Figure 233 (Richtigstellung)) addresses this problem by peremptorily merging the TP's duplicated Maeander valley routes into one. Geographically this might be comprehensible since all the stops on an ancient road proceeding about 100 Roman miles on the route Hierapolis - Laodicea - Trallis - Magnesia - Ephesus can be given. Graphically however, the Miller emendation is most implausible. Zigzag routes down the page are not part of the TP's visual code. I decidedly reject this and prefer to preserve the separation as drawn, even with the inconvenience of one road petering out at Tralles.

 At Magnesia, the TP linework has no hook, so I would argue it has been drawn not as a way-station but as a junction, with the route down to Ephesus omitted by the copyist. Talbert (TPPlace2229) argues from the sloping distance figure, XXX, that this line indeed existed and should be placed between the figure and the river. French did not research this route and gives no mention to the published milestone which Miller (725) alludes to.