The Aqueducts of Ephesos

Since 2001, the study of the Ephesian water supply has had as its goal the investigation, documentation, and publication of all of the aqueducts running into the city. In 2004, the 12th international symposium »Cura Aquarum in Ephesus« was organised in order to situate the investigation of the Ephesian aqueducts within the framework of international discussion, after the preserved inventory of all aqueducts had been previously reviewed and examined on the basis of publications up until now. During this process, sections of conduits, aqueduct bridges, tunnels, supply pipes and even two unknown aqueducts were discovered, so that now with this new information a more complex picture of the aqueduct system at Ephesos emerges. This picture extends

  • from the oldest aqueduct dating to the period of Lysimachos (first half of the 3rd century B.C.), the single earthenware pipe line of which is not longer than 1 km;
  • to the 8 km long Aqua Throessitica dating to the 2nd century B.C., to which a new line running over a monumental bridge was added shortly after the turn of the eras by Sextilius Pollio;
  • and to the late Hellenistic/early Roman Şirince Aqueduct, which was reused in the 6th century A.D., in order to supply the settlement which grew up near the Basilica of St. John with water by means of a pressure pipe with a 656 m long valley bridge at the end; and to
  • the Sultaniye-Aqueduct dating to the Augustan period with three clay ware pipe lines – probably the Aqua Iulia attested via inscriptions and most likely the predecessor of the Değirmendere Aqueduct; up to
  • the first channel conduit, the Trajanic Aristion Aqueduct, a long-distance conduit 210 Stadia in length coming from the Caystros Valley; this was led around the Mausoleum of Belevi in an artificial rock channel and ran under the steps of the Stadium and the Theatre until its provisional end point, the Nymphaeum Traiani on the Curetes Street; and finally
  • to the youngest (mid-2nd century A.D.) and longest (40 km) long-distance aqueducts, the Değirmendere Aqueduct, which ran over 22 bridges and 3 tunnels in a Qanat system of construction on the rear face of the large city hill, where it entered the city in the area of Tower 16 of the Lysimachean walls.


Since 2006, research has concentrated on the Değirmendere Aqueduct, which is the first of the long-distance aqueducts to be studied in detail, since due to the expansion of the city of Kuşadası this conduit in particular is endangered. Since 2009, it has been possible to intensify the investigations thanks to the financial support of the Austrian Science Fund (Project no. P20034-G02), with the result that the project is affiliated with the Kadikalesı Excavations of Ege University Izmir in Kuşadası.


The Değirmendere Aqueduct is divided into three areas:

  • In the first section, from the source up until Kuşadası, the conduit mainly is conducted over flat land underground, so that with the exception of two shorter sections only the bridges and two tunnels are known. Until twenty years ago, this section was still used for the water supply of Kuşadası. Due to the poor subsoil – the conduit here was built into a series of clay layers deposited in the sea – the bridges had to be renewed at a later date, especially as the Roman constructions were to a great extent washed away by rain water. Therefore, for example, one can plainly see at the Başkemer Bridge that it was renewed in the Genovese period (11th century A.D.) over meagre Roman remains, in order to supply the newly founded fortification of Scala Nuova with water. An additional building phase dates to the early 17th century, when once again one had to direct water to Kuşadası for the caravanserai erected by Öküz Mehmet Paşa; a final building phase dates to the 19th century. The Genovese conduit leaves the Roman aqueduct at the end of the Kalafat Tunnel, and is led into the city with a substantially smaller diameter over a new bridge, which is still preserved in its entirety in the suburbs of Kuşadası; in the built-up residential areas of the city today, however, no traces of the two conduits are preserved.

    Only the Zincirlikuyu-, the Kocakelle- and the Sabancık-Bridges, which were not eliminated during the extension of the modern city, can be identified as the original Roman bridges based on their design and masonry construction. During observation of the profile height of the Değirmendere Aqueduct it was apparent that the entire level of the neogene layers in the area beyond the Kalafat Tunnel up to the destruction line must have uniformly sunk to a maximum of 8 m, due to underground earthquakes, without resulting in serious damage to the bridges mentioned above – this phenomenon has also been confirmed by geologists.

  • Immediately beyond the last houses of Kuşadası, the second section begins; this extends to the Mezarlık Tunnel and in 2010, at least in the region of the Province of Aydın, it was possible to clear, clean and document a stretch of 10 km. Here the conduit is preserved high above the coastline and extends far into the Bahçecıkboğaz Valley. In the steep landscape, often supporting walls of up to 4 m were required. In the entire section, two conduits are present: the first part of the double conduit begins immediately to the north of the last settlement of Kuşadası up to the Bahçecıkboğaz-Bridge. Here, a new construction of the water conduit was necessary, since during an earthquake, due to sinking of the neogene clay layers, on one line the levels of the base of the duct suddenly gaped apart by 3 m. A new conduit with substantially smaller incline had to be erected, so that this could again flow into the old conduit near the Bahçecıkboğaz-Bridge. On the south side of the valley, the newer conduit ran above the older one, and in the direction of flow the difference of the base of the ducts becomes constantly smaller. Due to the extremely small incline in this section, the conduit was very vulnerable to interruption and frequently had to be repaired. Evidence of this is provided by six bypasses discovered until now, whereby the conduit was generally erected above the abandoned conduit and often – in order to obtain building material – by breaking its cheek on the valley side. It is possible that a mistake in surveying before the Mezarlık Tunnel, and also the decision to bring more water into the city, made necessary the new construction of the conduit in this section. To what extent shifting in the landscape played a role can only first be determined after an evaluation of the levelling. The discovery of six bypasses is an indication of the great difficulties in construction in this section. With these bypasses, the channel was mostly erected above the abandoned channel and also frequently – in order to gain more construction material – by the breaking up of its side wall on the valley side. This is rather sensational, as here, for the first time, the diversions during the course of repairwork to the conduit bed mentioned by Sextus Julius Frontinus in his 1st century A.D. work »De aqueductu urbis Romae« could be attested at a Roman aqueduct. Many supporting piers which were set up later on the valley-side wall of the conduit also provide evidence of the necessity for repair work on the aqueduct in this section; in section VIII.A, for a stretch of ca. 2 km, 123 of such piers were documented. In 2011 it was possible to discover that near the Bahçecıkboğaz-Bridge it was a case of a venter (viaduct) of a siphon; thus this highly sensitive structure constitutes the point of intersection of the two conduits. Since beyond this bridge the double conduit runs precisely inversely: here, the more recent conduit lies above the older one, and the difference in the base of the duct becomes constantly greater in the direction of the water flow. Due to the fact that it was also possible to discover the route of the older conduit along the city wall of Ephesos, thereby lying 8 m below the newer conduit, it is now clear that the second part of the double conduit was built for another reason: since the aqueduct was out of commission between the site of the disruption and the Bahçecıkboğaz-Bridge during the repair work, the Ephesians decided to construct a new conduit from the Bahçecıkboğaz-Bridge up to the city with a greater cross-section and higher level, so that areas of the city which lay at high levels could be supplied with the water from this aqueduct.
  • In the third section, from the Mezarlık Tunnel until the city, the aqueduct again mostly is underground, so that only the remains of bridges are visible. In 2007 and 2011, in sondage excavations, it was possible here to bring to light two sedimentation- and plunge-basins directly before and after the Arvalya Bridge, simple canal cross-sections, and the course beneath the Hellenistic city wall tower by which the aqueduct arrived in the city region of Ephesos. With the exception of the location at the Arvalyaçeşme Bridge and in front of the city walls, no traces of the old water channel are preserved due to the deep position; the old sections were probably removed from the plain during agricultural development. Only a densely overgrown spur of land below the Pisidere Bridge might conceal remains of the older bridge.



Duration of Project




Austrian Science Fund FWF




Bibliography (selected)

  • W. Alzinger, Beispiele antiker Wasserversorgungsanlagen: Ephesos, in: Frontinus-Gesellschaft (Hrsg.), Die Wasserversorgung antiker Städte II (Mainz 1987) 180–184.
  • P. Forchheimer, Wasserleitungen, in: FiE 3 (Wien 1923) 224–255.
  • Ü. Öziş – A. Atalay, Fernwasserleitungen von Ephesos, in: H. Friesinger  – F. Krinzinger (Hrsg.), 100 Jahre Österreichische Forschungen in Ephesos. Akten des Symposiums Wien 1995, AForsch1 = DenkschrWien 260 (Wien 1999) 405–411.
  • G. Wiplinger, Stand der Erforschung der Wasserversorgung in Ephesos/Türkei, Schriftenreihe der Frontinus-Gesellschaft 27, April 2006, 15–48.
  • G. Wiplinger, Der lysimachische Aquädukt von Ephesos und weitere Neuentdeckungen von 2005, in: Schriftenreihe der Frontinus-Gesellschaft 27, April 2006, 121–126.
  • G. Wiplinger, Wasser für Ephesos, Stand der Erforschung der Wasserversorgung, in: G. Wiplinger (Hrsg.), Cura Aquarum in Ephesus, Proceedings of the 12th International Congress on the History of Water Management and Hydraulic Engineering in the Mediterranean Region, BABesch Suppl. 12 = SoSchrÖAI 42 (Leiden 2006) 23–39.
  • G. Wiplinger, Neue Ergebnisse zur Wasserversorgung in Ephesos, in: Cura Aquarum in Jordanien, Proceedings of the 13th international Conference on the History of Water Management and Hydraulic Engineering in the Mediterranean Region, Petra/Amman 31 March – 9 April 2007, Schriften der Deutschen Wasserhistorischen Gesellschaft 12 (Siegburg 2008) 313–327.
  • G. Wiplinger, Die Wasserversorgung von Ephesos in byzantinischer Zeit, in: F. Daim – J. Drauschke (Hrsg.), Byzanz – das Römerreich im Mittelalter, Teil 2, 2 Schauplätze, Monographien des RGZM 84, 2, 2 (Mainz 2010) 593–613.




Gilbert Wiplinger