Alba Regia. Annales Musei Stephani Regis. – Alba Regia. Az István Király Múzeum Évkönyve. 25. – Szent István Király Múzeum közleményei: C sorozat (1995)

Tanulmányok – Abhandlungen - Christie, N.: The Survival of Roman Settlement alog the Middle Danube: Pannonia from The tenth Century A. D. p. 303–319. t. XX–XXIII.

oratories and even churches, as at Csopak, Fenékpuszta. Kékkút, Sümeg and Donnerskirchen. In cases like Kékkút and Sümeg the size of the basilical structure hints at an influx of people, perhaps away from the limes, and may indicate the provision of a chorepiscopus at the site (THOMAS 1964, 391-393; 1982, 266-271). Indeed, overall there are clear signs of major rural prosperity, comparable to villa booms in Gaul, Spain and Britain at this date and reflected primarily in the insertion of mosaics and the construction of granaries (horrea). The large, presumed imperial, villas such as Parndorf are the most opulent, examples, while the suburban villas I and II at Gorsium give an idea of general embellishment: in the case of villa II construction began only in the early 4th century, coinciding with Gorsium's rebirth; coins extend upto 375, and subsequently a cemetery formed around and part across the villa ruins (Bíró 1974, 39, 46, 52; MÓCSY 1974, 297-301; Fitz 1960; Bánki 1963-64). Economically Pannónia was on its feet, backed up by an agricultural exploitation previously lacking. We should suspect that this output was primarily geared to cater for troops not only along the limes but also for units of the new mobile army; simultaneously, villa prosperity suggests a shift of the wealth into the countryside. Significant is the emergence in the period 325-350 of a group of large, planned fortified complexes rearward of the Danube but strategically sited along the main roads. Five such centres are known, three of them (Heténypuszta/Alsóheténypuszta, Ságvár and Környe) in Valeria on the Sopianae-Brigetio road, and the remaining two in Pannónia Prima (Kisárpás on the Arrabona-Savaria road, and Keszthely-Fenékpuszta at the SW end of lake Balaton, between Savaria and Sopianae) {figs. 1 and 2). Although long known, only Ságvár and Fenékpuszta have been excavated to any degree. Each seems to have replaced an earlier Roman settlement located nearby: e.g. Ságvár for Tricciana and Kisárpás for Mursella; Heténypuszta has been equated with Iovia, whose name can be linked with the Tetrarchy. In plan the sites are roughly square, girded with large circuit walls reinforced by projecting round towers and featuring, two gates; internal structures, where studied, comprise a large wellbuilt granary, a pair of churches, a villa-like complex, and various smaller buildings (cf. fig. J); due to lack of publication, phasing is not always clear, but the granary and villa buildings appear to be primary features. In the case of Heténypuszta and Ságvár, fan-shaped towers underlie the round corner towers and may denote a Constantinian or Constantian origin (Barkóczi-Salamon 1984, 168-169; Tóth 1985, 123-126; cf. MÓCSY 1974, 304-306); the reinforcement of the walls and the construction of round towers (strikingly absent from the forts of the limes) is set to post-350 or even post-380. In effect, Fenékpuszta and Környe, which lack the fanshaped precursors, are viewed as late additions to this network of inner fortifications. Their role is much disputed: Mócsy views them as major imperial latifundia estates, dominating the rich agricultural land and acting as supply bases for the annona; he points out that the fortification of villas in the 4th century was only exceptionally permitted, hence the conclusion of imperial ownership (1974, 303-307; 1976). Further, he notes the possibility that tHeir defences could have been designed to intimidate the barbarians settled as federates on Pannonian soil. Soproni argues for a more specific military role as a rearward guard to the limes, controlling the main inland communications lines towards Italy: he stresses the size of the sites (from 500 x 450m at Heténypuszta, 380 x 360m at Fenékpuszta, descending to с 200 x 150m for Környe), each much larger than contemporary Danubian forts and with plans that smack of the work of military engineers and labour; the walls at Heténypuszta and Fenékpuszta both attain a thickness of 2.6m. The sites are not named in the Notifia Dignitatum, although its records for later 4th century Pannónia defy full interpretation. Nor need the lack of clear military structures within the fortifications argue against a military role, since late Roman forts in general rarely offer coherent internal plans: store buildings, granaries and churches do occur in stone, but barracks tend to be timber-built. The structures observed at Fenékpuszta are not therefore alien to a military fort (SOPRONI 1985, 1721; Tóth 1985, 126-128 on Ságvár; cd. JOHNSON 1983, 52-54, 123-124). The churches suggest that the sites may also have served the local rural population, providing refuge in times of crisis, and collection points for grain supplies (compare again details in the Vita Severini). Without excavation we cannot say if there was an overlap between these fortifications and the nearby Roman sites but the possibility is strong: Tóth's excavations at Ságvár, for example, revealed a secondary fortification phase which used spolia from the nearby Roman settlement; he offers the date of 374, coinciding perhaps too neatly with a major Quadic-Sarmatian invasion of Pannónia and the destruction of the horreum within the fortress. Subsequently, he argues, the refugee civilian population moved into the fortification and burials began up against the circuit walls (Tóth 1985, 124-127). Burials outside the south gate at Fenékpuszta likewise commence in the later 4th century, and extend into the 9th century (Barkóczi 1968, 275-277; Barkóczi-Salamon 1984, 181; Müller 1987). Throughout, however, there is a striking preponderance of civilian graves, with only occasional weapon burials - again suggestive of the idea of refuge centres. The Decline and Fall of Roman Pannónia (A.D. 364-430) Along the limes a new round of reinforcement took place after 350 in response to renewed Sarmatian and 305