This paper provides the threads connecting the three cities which form the basis of the three other papers in the panel. The paper is in two parts, which will be brought together at the end: the first considers the west-east routes coming from Sivas, sometimes Tokat, to Erzurum, and the second the routes coming up from Amid and Harput which cross the Antitaurus range and intersect with the west-east ones at critical points. The study of the west-east routes is essentially a reworking of the Ottoman itineraries in Taeschner's Wegenetz, though other itineraries are considered. However the whole argument is based, unlike Taeschner's, on the scheme of the network of built Roman roads which extended as far east as Erzincan. Until the mid-17th century all traffic from Sivas came through Erzincan. All traffic, in the westerly stages of the routes, took an apparent detour to the north through the Sushehri plain; but this detour both gave access to excellent pastures and avoided the steep inclines required by the direct line. At the Refahiye plain the routes diverged. Commercial traffic tended to take more southerly alignments, most easily to Kemah and then upstream along the Upper Euphrates valley. But the Ottoman armies of the 16th century took more northerly alignments, at a greater height, through the Chimen Dagh. Each of these alignments followed Roman roads whose lines seem to have been chosen so as always to allow stops at pastures. The principal south-north route crossed the Taurus range from Amid to Harput, then continued over the plain and crossed the Lower Euphrates. It then took advantage of a Roman road which ascended the Antitaurus via Arapkir. This road rejoined the banks of the Upper Euphrates and arrived at Kemah. In the period the evidence is for commercial rather than military movement along this line.