Savaria - A Vas Megyei Múzeumok értesítője 30. (2006) (Szombathely, 2007)

Régészet - Anderkó Krisztián: Savaria vízvezetéke

Anderkó Krisztián: Savaria vízvezetéke (Fig. 8: 18, Fig. 12). These technical solutions can be observed in respect to this section only, whilst other known cross-sections well confirm that such terrazzo was structured sector-like quateround in the internal corners and that the aquifer layer was directly situated on the substructure. Although, the channel was removed down to its inner bottom line (Fig. 13), the aqueduct recovered in 2001 must have been produced by the latter solution, since the terrazzo was also placed on the substructure in this case. The differences may refer to the chronological deviations, for which the reasons vary. As regards the aquaeductus in Köln, a certain part of the channel must have been produced to perform temporary duties (GMWE 1983: 351, Fig. 6; GREWE 1985: 40). However, this option is opposed by the fact that all of the sections recovered there were produced on a high quality level apparently for long-term use, whilst the supplementing section in Köln was made on a lower technical level in a smaller size as a temporary solution. It is also another option that the section recovered in the U3 was the result of a section modification taking place later, but this is less presumable owing to the expansive knowledge of the area and the track constrained by the topographical conditions. Whether the modification of the channel clearance may incidentally correlate to the service connection channel from a later period is still questionable; if we assume that the construction and the modification took place at the same time, and then another issue arises, namely why do the formations of the two aquifer layers differ? The most probable alternative for this is the full or partial reconstruction in respect to the channel clearance only, which was required due to hydrological reasons, perhaps. The air-raid shelters built in Street Kelemen Kárpáti in Szombathely have been identified with the large storage cellars marked on the cadastral map of 1965 having had a scale of 10.000 (Fig. 14: 33). Further issues arise in regard to the rest of the track and the ending point. The geographical locations of the Bagolyvár ('Owls' Castle) as a presumed castellum divisorium (Fig. 14:34) and the city surrounded by a wall compared to each other represent the problem. In most cases, this sort of facility was built on the highest point (HODGE 1992:280) within the city walls in accordance with the Vitruvius principle (VITRUVIUS 8. 6. 1.) and assured planned water distribution either through three water nozzles as defined by the Roman writer, such as in case of the castellum at the Vesuvius gate of Pompeii (GREWE 1985: 96), or through several more ones that were adapted to the local demands, for instance at least through ten in case of Nîmes (GREWE 1985: 95). Water was directed through lead, ceramics, wood or cement pipes from the water storage tank to the public wells, baths and private homes. The assumed end point represents a distance of 1.5 kms stretching between the Bagolyvár ('Owls' Castle) and the city center of Savaria surrounded by walls, which seems far too large for that laid pipeline network would have been applied to overlap it. If the Bagolyvár ('Owls' Castle) is only the end point of the subsurface section and the starting point of the vault section, the application of the term o/castellum divisorium is incorrect. At the same time, the construction change itself would not have demanded the installation of the premise described as having had a width sufficient for 'two wagons' into the track (TÜRR 1953: 132), except for that if this had not been justified by reasons related to hydrology (pressure decline chamber

— GREWE 1985: 77) or water cleanness (settning tank — GREWE 1985: 85—91). Examining this issue from a different point of view, we see that in case of the subterranean channels found in the Köln and Trier vaults are installed in those cases, where the distance between the city and such subterranean channel is the shortest (GREWE 1983: 344, Fig. 1; NEYSES 1980: 4, Fig. 1). Therefore, the maximal exploitation of the opportunities provided by the 42