Introduction

The principal visible remains at Yassıhöyük (Gordion), Ankara province, Turkey, are its large flat-topped höyük, the 'Citadel Mound' (Figure 1) and some 100 tumuli, several of exceptional size. These earthworks were 'discovered' in 1893 when building the Berlin-Baghdad railway, and the location subsequently identified as the Phrygian capital of Gordion (Reference SamsSams 2005: 10). Between 1950-73, Gordion was the focus of excavations by the University of Pennsylvania Museum aimed at bringing 'archaeological light' to this obscure Iron Age society (Reference SamsSams 2005: 12). In the process, evidence was recovered for extensive Roman (and later) activity on the Citadel Mound but little consideration was then given to establishing its precise nature. However, a critical re-assessment of all periods of occupation identified in the 1950-73 excavations at Gordion took place when a new research phase commenced there in 1988. This analysis revealed inter alia that Roman activity on the Citadel Mound began after Provincia Galatiae was formed in 25-20 BC; reached a climax in the Flavio-Trajanic period; and sharply declined with a brief renaissance c. AD 300 (cf. Reference VoigtVoigt 1994 with Reference GoldmanGoldman 2005: 59-60, 66).

Figure 2.

'Snake's-head foot' brooch of Central European origin.

Re-discovering Roman Gordion

The 1988 analysis strengthened the supposition that Gordion was the Vindia named in Roman geographical sources, and inspired a comprehensive review of the Roman-period material recorded at the site (Reference GoldmanGoldman 2000). This review demonstrated that certain of the Roman structures at Gordion represented a planned settlement of probable Flavio-Trajanic date (Reference GoldmanGoldman 2005: 63, 66), and also isolated a range of associated ceramic forms without clear parallels in Anatolia, although subsequent research identified comparable material on Julio-Claudian and Flavio-Trajanic military sites along the Rhine and Danube (e.g. Gassner & Jilek 1997: Figure 5.15). Moreover, a Roman military tombstone of Trajanic date was found north of the Citadel Mound in 1996 (Reference GoldmanGoldman 1997), while further research established that three 'Roman' brooches found at Gordion between 1950-1966 were of types common to first-century AD military contexts in Central Europe (e.g. Figure 2), raising the possibility that the early Roman settlement at Gordion included a military element.

In order to establish more positive information regarding the nature of the early Roman settlement at Gordion, four 9 x 4m. trenches were opened on the Citadel Mound in 2004, with further work on these in 2005. Each trench provided artefactual evidence for a Roman military presence at Gordion, the most conclusive evidence being found in our Operations 54 and 55. Briefly stated, these uncovered a road surface immediately west of a 9m wide structure of unknown length and of mixed masonry and mud-brick construction (Figure 3). The building had at least three contiguous rooms on its linear axis, with the central one at least divided by a cross-wall to create a smaller western front space and a larger rear area. This plan corresponds to that of a contubernium, the 'living space' found in Roman military barracks of imperial date, with the arma ('armoury') to the west and the papilio ('sleeping area') to the east. Thus, as the rooms on either side formed part of the same construction, all three rooms probably represent contiguous contubernia in such a building.

Confirmation that this structure was probably a Roman army barracks was provided by the artefactual material indirectly and directly associated with it. This included copper alloy harness pendants of Roman military type, one closely paralleled in Roman military contexts in the Balkans (Figure 4; cf. Reference Radman-livajaRadman-Livaja 2004: 480), while a midden 'behind' the building, and provisionally dated by glassware to the late Julio-Claudian or Flavian period, produced individual examples and linked groups of iron rings from ring-mail armour and single examples of copper alloy and iron scales from scale armours. Even more conclusive was the discovery of two lengths of flexible scale armour, the iron scales joined with copper alloy staples (Figure 5), in what would have been the arma of one of these assumed contubernia.

Figure 3.

Early Roman features in Operations 54 and 55.

Figure 4.

Military harness pendant of Balkan type.

Figure 5.

Military harness pendant of Balkan type.

Discussion

Analysis of the 2004 and 2005 work at Gordion continues, but there can be little doubt that Operations 54 and 55 revealed a Roman army barrack block. Gordion thus becomes the first site in Turkey to provide an example of a standard building form exclusive to regular Roman army units of the imperial period. The glassware found with it suggests use from possibly the late Julio-Claudian or early Flavian period, while some of the pottery and metalwork has Central European parallels: in which case literary evidence for Roman troops being sent from that region to assist in Nero's Armenian campaigns of 57-61 may well be of relevance (Tac. Ann. 13.35, 15.6, 15.25; cf. Bennett, in press). It is not immediately clear why a garrison was needed at Gordion in the mid first century, but the place may have served as a focal point for assembling food stocks for onward despatch during the Armenian campaign. If so, then Gordion could have remained in use as a supplies base until the end of Trajan's Parthian War (the date of the tombstone), only to became redundant when Hadrian rationalised army deployment in Anatolia in 117 (Reference BennettBennett 2007: 139).