Pompey destroyed most of the castles, supposedly because they could become refuges for robbers and brigands, in reality probably because in the hands of disloyal local dynasts, they could be a threat to Roman control.
The administrative units in Pontos seem to have been quite small. We know from the inscriptions in the sanctuary of Zeus Stratios that the territory of Amaseia was divided into at least twelve districts, only five of which are mentioned by Strabon.6 By the second century AD, when the inscriptions were erected, these districts had long ceased to have any administrative function, but their names lingered on. Almost all the districts have indigenous names ending in -ηνη or -ιτις, probably with roots going back to the Bronze Age, and the same is true for the villages or hamlets mentioned.7 On account of the inscriptions, D. French suggests that there may have been as many as 500 villages in the territory of Amaseia,8 which of course brings to mind the district northwest of Amaseia called Chiliokômon, the “valley of a thousand villages”. Contrary to most names of districts, this carries a Greek name, but it could of course be a Greek translation of an earlier, indigenous name. All this suggests a densely populated rural landscape in the valleys of the interior during the pre-Roman period. Supposedly it was from all these scattered villages that the inhabitants of the Roman foundations were drawn.
There are, however, reasons for not accepting this interpretation of pre- Roman decentralisation contra Roman urbanisation too readily. Over the past few years, our knowledge about settlement patterns and land-use in central northern Asia Minor has increased significantly, mainly due to two international survey projects carried out in Paphlagonia in the late 1990s. Many other, less intensive surveys have been carried out in Pontos by local archaeologists.
Fig. 1. Satellite image of the Sinop Promontory.
The Sinop Survey
During the late 1990s, the Sinop Regional Archaeological Project intensively surveyed 350 selected tracts on the Sinop promontory and located 170 archaeological sites or loci (Fig. 1). The preliminary publications of the survey have demonstrated that until about the time when Sinope became the capitol of the Pontic kingdom, habitation was scarce on the Sinop promontory, and the city’s primary lines of communication went via the sea.9 The site density in most areas is very low in the archaic and classical periods, with a slight rise during the Hellenistic period, but the real change in the settlement pattern only occurs under the Empire. Exceptions are the smaller promontory Boz Tepe (which, due to its geographical position, had always been closely linked to the city) and the area around Armene west of Sinop, where Xenophon and the Ten Thousand made landfall on their westward journey by sea along the coast. The director of the project, Owen Doonan, concludes that in contrast to the earlier periods, “Roman settlement tended to be extensive, specialized and connected”.10 The change is particularly noticeable in the Demirci Valley to the south of the city and the Karasu Valley to the north (Fig. 2), which saw the growth of an extensive olive production and the production of amphorae in the second century AD.11 Smaller hamlets, probably relying heavily on marine resources, also began to dot the coastline during this period, and suburban elite settlements such as the possible villa site at Kiraztepe were established.12
Fig. 2. The number and size of loci in different areas of the Sinop Promontory in the Hellenistic (top) and Roman (bottom) periods (from Doonan 2004, 156-157).
The Paphlagonia Survey
Likewise in the late 1990s, a British team carried out a somewhat more extensive survey of the towns of Hadrianopolis and Antoninopolis in the interior of Paphlagonia.13 Their conclusion about the development of the settlement pattern was similar to that of the Sinop peninsula survey: “During the Roman period, settlement in southern Paphlagonia takes on new dimensions. For the first time we start to see widespread settlement across almost the entire landscape … and the first appearance of a truly distinct hierarchy of settlement, ranging from large town to small hamlet with associated cemetery”, and further: “A notable component of the Roman settlement pattern is the prevalence of small sites”.14 Precise dates are not given for specific sites, but coins and inscriptions of the third century AD are mentioned. In all, thirtyone loci occupied in the Roman period were identified. Only seven of these loci (23%) also contained material from the Hellenistic period. Moreover, in no instances does occupation terminate during the Hellenistic period. The complete site continuity from the Hellenistic to the Roman period indicates that in the later period, inhabitants spread over the landscape from already existing sites.
The reports from both surveys point to the peaceful situation during the Roman imperial period as the primary reason for the characteristically dispersed settlement pattern, which was not matched until modern times.
Other surveys
Several surveys have been carried out further eastward in Pontos by Turkish archaeologists. Projects initiated by M. and N. Öszait cover the districts of Amasya, Samsun and Ordu, and another group has been working in the areas around Tokat. In Paphlagonia, a team has been working in the area around Kasamonu.15 However, none of these surveys follow as systematic an approach as that mentioned above, nor are they as intensive. In addition, most of the projects focus predominantly on periods earlier or later than the Hellenistic and Roman era. What can be glimpsed from the many preliminary reports boils down to a generally wider distribution of sites in the Roman period. The evidence will not at present support broader conclusions because of the preliminary nature of many of the publications, and because the Hellenistic period was shorter than the Roman period – which is often taken to include the Byzantine period as well – and also because in extensive surveys Roman remains may be more readily recognizable than Hellenistic ones.
The sum of available evidence creates an impression of a Roman landscape that, contrary to what might be expected, did not concentrate settlement in the urban centres that were the focus of the Roman administration, but rather distributed the population across the countryside in a variety of settlement types. This development under Roman rule is quite different from that of the province of Achaea, by far the most thoroughly investigated area, where “the Classical and early Hellenistic periods appear exceptionally active, characterized by the presence of numerous, dispersed, small sites”. By comparison, with few exceptions the Roman landscape of Achaea appears “empty”.16 However, Achaea was probably not representative of the Empire in this respect, and developments in northern Asia Minor are parallelled elsewhere. This raises a question about the demographic potential of Pontos during the Hellenistic period. Where did Mithridates obtain soldiers for his campaigns against Rome if there were no large urban centres and little dispersed settlement?
Defining time – the use of calendar systems in northern Asia Minor
Space was altered as new settlement patterns changed the landscape and the administrative infrastructure was shaped along new lines when the land was parcelled out among newly created cities. Time, or rather the reckoning of time, changed as well.
Prior to the Roman conquest, the dominant system for reckoning years in northern Asia Minor was according to the Bithyno-Pontic era, counting the years from the accession of King Zipoites of Bithynia in 297/96 BC.17 The earliest evidence for its use are coins struck in 149/48 BC, the year Nikomedes II became king after the murder
