ChapterPDF Available

The Hadrianic aqueduct of Athens and the underlying tradition of hydraulic engineering

February 2018

DOI:10.2307/j.ctvndv54v.9

In book: Great Waterworks in Roman Greece (pp.70-97)

Authors:

Eustathios Chiotis

Institute of Geology and Mineral Exploration

Content uploaded by Eustathios Chiotis

Content may be subject to copyright.

Great Waterworks in

Roman Greece

Aqueducts and Monumental

Fountain Structures

Function in Context

edited by

Georgia A. Aristodemou and

Theodosios P. Tassios

Archaeopress Roman Archaeology 35

Archaeopress Publishing Ltd

Gordon House

276 Banbury Road

Oxford OX2 7ED

www.archaeopress.com

ISBN 978 1 78491 764 7

ISBN 978 1 78491 765 4 (e-Pdf)

Cover: The monumental arcade bridge of Moria, Lesvos,

courtesy of Dr Yannis Kourtzellis

Creative idea of Tasos Lekkas (Graphics and Web Designer, International Hellenic University)

by any means, electronic, mechanical, photocopying or otherwise,

without the prior written permission of the copyright owners.

Printed in England by Oxuniprint, Oxford

This book is available direct from Archaeopress or from our website www.archaeopress.com

Contents

Preface ��iii

Georgia A. Aristodemou and Theodosios P. Tassios

Introduction I� Roman Aqueducts in Greece ��1

Theodosios P. Tassios

Introduction II� Roman Monumental Fountains (Nymphaea) in Greece ��10

Georgia A. Aristodemou

PART I: AQUEDUCTS

Vaulted-roof aqueduct channels in Roman Macedonia ��15

Asimina Kaiafa-Saropoulou

The aqueduct of Actian Nicopolis ��26

Konstantinos L. Zachos and Leonidas Leontaris

The water supply of Roman Thessaloniki ��50

Manolis Manoledakis

The Hadrianic aqueduct of Athens and the underlying tradition of hydraulic engineering ��70

Eustathios D. Chiotis

The Hadrianic aqueduct in Corinth ��98

Yannis Lolos

The Roman aqueduct of Mytilene �� 109

Yannis Kourtzellis, Maria Pappa and George Kakes

Roman aqueduct of Samos �� 131

Τelauges Ν. Dimitriou

A Roman aqueduct through the Cretan highlands – securing the water supply for elevated Lyttos � 147

Amanda Kelly

PART II: NYMPHAEA

Shifting tides: approaches to the public water-displays of Roman Greece ��173

Dylan Kelby Rogers

Fountain figures from the Greek provinces: monumentality in fountain structures of Roman

Greece as revealed through their sculptural display programs and their patrons �� 193

Georgia Aristodemou

The monumental fountain in the Athenian Agora: reconstruction and interpretation��218

Shawna Leigh

New water from old spouts: the case of the Arsinoe fountain of Messene �� 235

Mario Trabucco della Torretta

Reflecting the past: the nymphaeum near the so-called Praetorium at Gortyn ��246

Brenda Longfellow

The Hadrianic aqueduct of Athens

and the underlying tradition of hydraulic engineering

Eustathios D. Chiotis

Αρχή παιδεύσεως ονομάτων επίσκεψις1

(The investigation of the meaning of terms is the beginning of education)

Antisthenes (445-360 BC)

2 to Theodosios Tassios

Abstract

The Hadrianic aqueduct of Athens is a monumental work of the Roman Emperors Hadrian and Antoninus Pius, constructed

during the period AD 125 – 140, with unique hydraulic features. The backbone of the aqueduct is an array of deep wells along

a 17.8km long continuous tunnel, whereas the typical Roman aqueducts of Greece (Nikopolis, Corinth, Samos, Lesbos, Knossos,

Gortyn) are mainly constructed on the ground and are characterized by water bridges upon arcades across valleys, although

occasionally comprise tunnels in order to pass through topographic highs.

In the dry land of Attica (of poor surface water resources), the Hadrianic aqueduct mainly captured groundwater along the

tunnel, whereas the typical Roman aqueducts of Greece transferred water from distant springs or rivers. The route and the

constructional details of the tunnel are adequately known, given that a section of the aqueduct was recognized and located in

1840 by Friedrich Stauffert3 and remained the main hydraulic work supplying Athens with water till the 1930s. The sustainability

of the aqueduct is due to its design for underground water capturing, a concept already applied to the archaic aqueducts of

Aegina and Megara seven centuries earlier.

Despite the fragmentary nature of archaeological data, a preliminary effort is undertaken to trace the remains of the water

distribution network in the city of Athens and correlate conduits, reservoirs, fountains and baths associated with the Hadrianic

aqueduct. Furthermore, its period of operation is examined, as well as the reasons for the construction of a Late Roman aqueduct

with water bridges, still preserved in Athens.

Keywords: aqueduct, groundwater, Hadrian, Athens, tunnel, water bridge

Introduction

The Roman Emperor Hadrian was a dedicated philhellene, especially fond of Athens and among his

numerous benefactions to the city was the aqueduct. Hadrian visited Greece three times when he was

Emperor (AD 125, 129 and 132) and, as commonly stated, the aqueduct was founded most likely during his

first Imperial visit (AD 125).

The main structure of the Hadrianic aqueduct is a 17.8km long tunnel, the northern end of which lies

in the center of the modern Olympic Village of the 2004 Olympic Games, at the foothills of the Parnes

Mountain, at a depth of 30.5m (Figure 1).

The tunnel ends near the Haghios Demetrios Church, at Ampelokipoi (AD in Figure 1), at a bottom depth

of about 6m.4 From this point it continued as a covered channel constructed in a few meters deep trench

for another 2.2km up to the Lycabettus Hill near the city of Athens. The water was conveyed by gravity

into a reservoir, in front of which a monumental façade was built. Thanks to early travelers, descriptions

As cited in Arrian (Discourses of Epictetus, i, 17).

(EN: Gift in return).

Papageorgiou-Venetas 2010: 109

Some critical elevations of the tunnel bottom are: at the well 299 in the northern end 212m, at Haghios Demetrios c. 140m and at the bottom

of the Lycabettus reservoir c. 133m. The elevation at the NE boundary of the National Garden is c.100m, high enough for the natural flow of water

to the ancient city of Athens.



Figure 1. The course of the Hadrianic aqueduct from the North (well 299 at Demogli) up to the resevoir on the Lycabettus Hill;

the surface channels at Demogli (1 and 2); the modern branches (3 and 4); the Acharnian aqueduct (from the Parnes Mt to

Athens); lateral modern extensions 6 to 9 and the ancient Chalandri branch 10 and 11.

The Hadrianic aqueduct of Athens

Eustathios D. Chiotis

Great Waterworks in Roman Greece

and drawings of the façade and its inscription (according to which the aqueduct was completed in AD

140) are available to us. A fragment of the inscription can still be seen in the National Garden of Athens.

Every 35-40m, wells up to 41m deep were dug along the tunnel using the shafts-and-gallery technique.5

The aqueduct fell into obscurity during Late Antiquity, apparently due to tunnel collapse near the Haghios

Demetrios church at Ampelokipoi. The water continued to overflow and was therefore considered as a

spring. In the Ottoman period this ‘spring’ was exploited for irrigation of the local fields and the left over

was conveyed to Athens via a stone-built channel. Experts knew that the Haghios Demetrios ‘spring’

was associated with the Hadrianic aqueduct, the aligned wells of which were still visibly full of water.

Friedrich Stauffert, an architect of the Municipality of Athens during the period 1835-1843, when Athens

was still a small town of 25,000 people, tracked the aqueduct for about 5km up to Chalandri, where he

lost the continuation.6 In 1840, he counted 109 shafts of an average section of 1.5m2 and was the first to

recommend the cleaning of the aqueduct for the water supply of Athens.

The process of i) tracing the tunnel ii) cleaning the aqueduct from mud and collapsed blocks of rock and

iii) supporting the walls of the tunnel was an expensive and long project, which started in the middle of

the 19th century and took a few decades to complete; 465 shafts had been reported cleaned by 1879.7 The

aqueduct remained the main water supply system of Athens until the Marathon Reservoir was completed

in 1935 – which for the first time supplied water from areas outside of the hydrologic basin of Athens,

surrounded by the Mountains Hymettus, Pentelikon and Parnes (Figure 1). During the summer period, the

Hadrianic aqueduct, at its maximum capacity, supplied modern Athens daily with ca. 10,000m3 (116lit/

sec) of water and was in operation until the middle 20th century. After a chemical analysis conducted in

1893, the quality of the water proved to be very good, due perhaps to CaCO3 deposition along its route. 8

In addition to the detailed description of the aqueduct by Kordellas, Ziller (1877) also wrote short remarks;

Pappas (1999) provided plans and data, but the first systematic study was accomplished by Susan Leigh

(1998), whose excellent treatise constitutes a dependable source, although important new evidence was

brought to light from recent excavations. Impressive underground observations and images have been

recently uploaded from a continuing research program conducted by the engineer and speleologist P.

Defteraios, in preparation for his Ph.D. dissertation.9

The sustainability and some constructional details of the aqueduct have been recently described;10 this

paper shall emphasize on other aspects and mainly on the tunneling technique many meters below the

water table, the similarities with the Raschpëtzer aqueduct in Luxemburg and the water distribution

network in Athens. In addition, the evolution of pre-Roman hydraulic technology in Greece and its

application to the Hadrianic aqueduct will also be examined and correlated to the Iranian qanats in the

Appendix.

Description of the aqueduct

The aqueduct tunnel collected surface and groundwater of various origins. Two recently excavated

surface channels at the foot of the Parnes Mountain conveyed water from springs in modern times:

Channel 1 was part of the so named Karydhia aqueduct,11 the northern extension of which is only loosely

known, and Channel 2, part of the so called Varybopi12 aqueduct which continues as a shaft-and-gallery

The manholes were sunk at regular intervals of 35-37m, according to Travlos (1971: 242).

Cited in

Kordellas 1879: 85.

Praktikatēs en Athēnais Archaiologikēs Hetaireias

water at the Ilissos riverbed is compared to that of the Hadrianic aqueduct; chemical analysis indicates that the water sample from the Lycabettus

Hill reservoir was of much better quality than from Kallirhoe (E. Ghintoni, IGME, oral communication).

http://urbanspeleology.blogspot.gr/p/blog-page_2754.httm

Chiotis and Marinos 2012: 23, 36.

Pappas 1999: 190.

Pappas 1999: 190.



aqueduct up to Haghia Kyriaki and further northwards (modern branch 2). At point A, Channels 1 and 2

are joined together into a single one which pours into the inclined gallery BC (Figure 2).

Pappas supported that the well 299 at C was the northernmost end of the ancient aqueduct and that all

hydraulic works northwards were modern extensions.13 This view was also adopted by the excavators of

the site.14 Asimakou and Kalliga, suggested in particular that the surface Channels 1 and 2 are works of

the last quarter of the 19th century.

Although the Karydhia and Varybopi aqueducts are modern extensions at the end of the 19th century, it

is supported here that the Channels 1 and 2 have replaced ancient ones, transferring water from springs

into the aqueduct, not directly to the well 299, but through the typical inclined Roman gallery BC, which

was constructed exclusively for this purpose (Figure 2).

Along the section AB it is possible to distinguish ancient remains: whereas the channels 1 and 2 were in

general covered in modern times with plates of Pentelikon marble, ancient stone plates of well cemented

breccia are maintained along the section AB, quarried locally in the ancient technique of opening a

narrow groove around the block to be excavated. Thus, the rock dug from the excavation of the trench

was used in antiquity to cover the trench. During the modern operation of the aqueduct, the ancient

trenches were extended and plastered.

The Hadrianic aqueduct was constructed almost 5 centuries after the so called Acharnian aqueduct (dated

in the 3rd quarter of the 4th century BC), the course of which is approximately traced in Figures 1 and 2,

based on the ancient inscriptions found along its course.15 The northernmost inscriptions (Figure 1, points

In1 and In2), were found 1500m north of the well 299. It would be therefore unwise for the engineers of

the Hadrianic aqueduct to construct the inclined gallery and let the adjacent springs (the main resources

of the Acharnian aqueduct) unexploited. It is therefore supported that the ancient channel 2 collected

water from these springs.

The inclined gallery, which connects the surface Channels 1 and 2 to the Hadrianic tunnel is 80m long, dips

ca. 20o. It is a typical Roman construction: the walls are made in opus latericium and the roof is composed

of three terracotta bricks, which are slightly rotated forming a pattern known as ‘pitched-brick’ vaulting

(Figure 3).

According to Lancaster, the earliest known Roman examples of ‘pitched-brick’ vaulting with curved bricks

appear in hydraulic systems in the 1st half of the 2nd century AD at Athens, Eleusis, and Argos, while

others, at Dion and Gortyn, may also date from the same period.16 It is furthermore argued by Lancaster,

that the technique was introduced from Parthia by military engineers, specialized in hydraulics, who

participated in Trajan’s Parthian War and, upon returning home, they brought along new construction

techniques.

This view should probably be reconsidered in the light of opposing evidence. Similar bricks have been

used in Eleusis at a grave dated in the middle 1st century BC.17 Furthermore, vaulting technique was used

in much earlier tunnels in Assyria, in the Palace of Sargon II at Khorsabad.18

An additional control point (In3) of the Acharnian aqueduct is provided by an inscription recently

excavated near the Hadrianic aqueduct reported at a depth of 8m and identified via the characteristic

feature of the elliptical section of the excavated conduit.19 However, the fact that it was found at the

The well 299 in Figure 1 corresponds to the point C in Figure 2.

Platonos-Yiota 2004: 61; Asimakou and Kalliga 2013: 167.

Vanderpool 1965: 171.

Lancaster 2010: 464.

Mylonas 1932: 45.

Mylonas 1932: 48.

Platonos-Yiota 2004: 440; Georgousopoulou 2013: 156.

Great Waterworks in Roman Greece

depth of 8m (the

Acharnian aqueduct

was usually placed

in a cut-and-covered

trench, only a couple

of meters deep), raises

the question whether

the conduit belonged

to a system supplying

water from the

Hadrianic aqueduct

to the Roman farms in

the area.

A boundary stone

(HOPOC) found at

Metamorphosis was

considered as a proof

that the Acharnian

aqueduct reached

this area.20 However,

according to the

excavation report, it

was found at a distance

from the course

of the Acharnian

aqueduct and exactly

on the route of the

Hadrianic aqueduct

(point HP in Figure

1). Thus, a ‘HOPOC’

stone without a more

specific indication

does not constitute a

solid evidence for its

association with the

Acharnian aqueduct.

Another section of the

Acharnian aqueduct is

suggested here for the

first time, at the point

where the aqueduct

crossed the ancient

city walls (Figure 1). A water duct was excavated in 1959 by Threpsiades, passing through an opening of

the ancient walls, of similar structure to the Acharnian aqueduct. 21 The terracotta U-shaped half-drums,

each of internal dimensions 0.42m height by 0.40m width, bear a combination of Greek letters, such as



excavator to the beginning of the 4th century BC and its abandonment period is dated by a late Roman

Platonos-Yiota 2004: 59 and 70.

Archaiologikon Deltion

Figure 2. Projection on Ktimatologio air photos: The Olympic Village and the surface

channels 1 and 2; the inclined gallery BC; the northernmost end of the Hadrianic

tunnel (C), and the Acharnian aqueduct; the locations of the Inscriptions 1 and 2 (In1

and In2).



Figure 3a. Top. The entrance

to the inclined gallery at the

Olympic Village. The walls

are made of opus latericium,

and the roof is composed of

three curved, slightly rotated

bricks. (Photo: Courtesy of P.

Defteraios).

Figure 3b. Bottom. The lower

part of the inclined gallery

at the Olympic Village. The

roof is composed of corbeled

rectangular bricks; further

down, by horizontal stone

plates (Photo: Courtesy of P.

Defteraios).

lamp found in the sediment filling the aqueduct. However, it is reasonable to assume that the Acharnian

aqueduct was abandoned earlier, when the Hadrianic aqueduct was put in operation.

The extension of the above described aqueduct was most likely revealed and retained in place in a recent

excavation, during the foundation of the new building of the Technical Chamber of Greece in Athens

(3-5, Palaion Patron Germanou Str.); however, the excavation report is not published yet. It is suggested

that the water of the Acharnian aqueduct from this point (at elevation of c. 72m) could possibly reach the

Agora of Athens by gravity flow. It could certainly supply Plato’s Academy as well, along the Demosion

Sema, as inferred from the Hellenistic pipeline at the site C8 in Figure 7 (described below) – which could

represent a branch of the Acharnian aqueduct at the site of the Kimonian branch of the Peisistratean

aqueduct.22

Chiotis and Chioti 2012: 429.

Great Waterworks in Roman Greece

Lateral underground branches of the Hadrianic aqueduct, approximately parallel to the course of the

main streams, supplied additional water. The better known and most important is the ancient Chalandri

branch which brought water from the Pentelikon Mountain.23 However, the most significant source of

water was and remains the groundwater along the wells and the main tunnel excavated below the water

table.

The tunnel bottom of the 30.5m deep well 299 lies 10m below the undisturbed water table,24 and remains

below the water table over a distance of approximately 800m (Figure 4a).25 Downstream, where the older

cone of compacted sediments occurs below a thin alluvial mantle, the wells are rather shallow, slightly

deeper than 10m and stop in the upper zone of the hard rock. The tunnel depth in general varies mostly

between 15 to 35m (Figure 4b) – except at the final section (Haghios Demetrios church), where it gradually

becomes shallower until it surfaces. Longitudinal sections of the Figure 4 were compiled from original

data kindly provided by the EYDAP Engineer E. Nestorides.26

The Hadrianic tunnel was intentionally designed unusually deep in order to tap groundwater and thanks

to that the aqueduct is sustainable and still functional. The water is infiltrated through both the wells

and the tunnel itself. The average gradient of the tunnel is ca. 0.4% although it is locally higher near the

Kifissos River, where the tunnel has to pass below the riverbed and the adjacent streams at a safe distance.

Pappas 1999: 187. Sections 10 and 11 in Figure 1 come from Kifissia Sheet (Curtius and Kaupert 1881).

This numbering starts from Haghios Demetrios. Earlier numbering started from the Lycabettus Reservoir and the northernmost well was

numbered 367, considerably fewer than the 465 wells reported by Kordellas, due to loss of wells. Assuming a 37.5m average distance between

wells, 475 wells are estimated along the proper tunnel, 17.8km long.

Chiotis and Marinos 2012: 27.

Chiotis and Marinos 2012: 31.

Figure 4a. Top. Section along the northern part of the Hadrianic aqueduct depicting: the ground surface, the water table, the

inclined gallery and the aqueduct (after Chiotis and Chioti 2012: 432).

Figure 4b. Bottom. Surface elevation vs tunnel mileage measured from Haghios Demetrios to the Kalyftaki stream (section

restricted by the available data. After Chiotis and Chioti 2012: 433).



Hydrogeological conditions

The Hadrianic aqueduct was built completely underground in order to benefit from the local

hydrogeological conditions. This deep underground design intended to capture groundwater through an

area of lower precipitation, lacking sufficient surface water resources.27

There is, however, a common view that the Perissos and Philothei water bridges belong to the Hadrianic

aqueduct (Figure 5). Leigh has convincingly shown that at both water bridges their elevations are lower

than the adjacent Hadrianic tunnel and could not therefore supply it with water.28 Most likely, the Perissos

and Philothei water bridges belong to a Late Roman aqueduct which collected water from springs at

Kifissia and Heracleion.29

The geological formations along the course of the aqueduct, shown in Figure 5, are:

•Quaternary talus sediments

•Older foothill cones

•Red Pikermian sediments

•Lignite-bearing Neogene marls and clays and

•The metamorphic basement represented by Athenian schists.

Thick Quaternary talus sediments outcrop near the Parnes Mountains and constitute the best aquifer in

the area. The older formations: Older Cones, Pikermian, Neogene and Schists are impermeable, covered

by an alluvial aquiferous mantle. The aqueduct taps water mainly from the Quaternary cone over a

length of c. 800m, by means of wells 20-30m deep and the tunnel itself (Figure 4a). Further downstream,

in the stratigraphically underlying older cone of compacted sediments, the wells are less deep. The whole

design of the aqueduct indicates a good knowledge of the hydrogeological conditions in advance which

could only be obtained from previously dug wells. The stratigraphy and water yield of such wells would

be decisive for planning the course and the depth of the aqueduct.

Copying the ancient engineers, the groundwater of the Quaternary breccias near the northern end of the

Hadrianic tunnel was also captured in 1901-1903 along a modern shafts-and-gallery extension (Modern

branch 1 in Figure 2) with 7 shafts 33m deep, connected through a tunnel 320m long. The project was

undertaken by Petros Protopapadakis, Chief Engineer of the Municipality of Athens (and later, Prime

Minister of Greece), and yielded an additional summer rate of about 750m3 daily.30

Along the aqueduct’s route additional water is collected from the perched aquifers on top of the above

described impermeable formations. If occasionally good aquifers are crossed by the wells, lateral galleries

are opened above the tunnel floor to tap them.

Remarks on tunneling

Tunneling faced two significant problems: rock stability and digging at rather significant depths below

the water table. The first one was solved by minimizing the tunnel dimensions and by constructing

artificial support. The tunnel dimensions were marginally ergonomic for a small in size person, i.e. 0.4-

0.8m wide and 1.2-1.8m high. The walls were supported by stone masonry, whereas the roof by either a

brick vault (Figure 3a) or a corbelled roof (made of stone or terracotta plates) (Figure 3b).

According to Ziller, the wells and the tunnel north of Ampelokipoi were dug in solid rock.31 The wells

are 9-15m deep and their diameter varies from 1.20m -1.50m. The tunnel measures 0.70m in width and

The annual precipitation is c. 400mm in Athens and almost doubles on the Parnes and Pentelikon Mountains.

Leigh 1998: 61-65.

Chiotis and Chioti 2011: 182, 188.

Paraskevopoulos 1907: 446.

Ziller 1877: 121-122.

Great Waterworks in Roman Greece

1.60m in height and the walls are covered with waterproofing plaster (perhaps to avoid water loss at

tunnel elevations above the water table). In case of fractured rock walls or soft formations the walls are

supported with masonry and the roof by a vault of three terracotta bricks.32 Beyond Koukouvaounes

(modern Metamorphosis), the tunnel crosses thick clay layers and the wells are lined with horizontal

layers of 6 curved bricks.33 Wall masonry and roof support by curved bricks, stone or terracotta plates are

also described in a recent publication.34

In addition to Haghios Demetrios, another location where the tunnel almost surfaces is at the southern

bank of the Kifissos River; this location is ideal for initiating part of the aqueduct by digging upstream.35

Uphill tunneling and water drainage at the excavation front would cause a local depression of the water

Ziller 1877: table VIII-28.

Ziller 1877: table VIII-6.

Angelakis et al. 2014: 100

At this region the Kifissos River could also supply the aqueduct, but there is no such evidence.

Figure 5. Projection on

ESRI’ world digital map:

geological formations along

the tunnel of the Hadrianic

aqueduct; Late Roman

water bridges at Philothei

and Perissos.



table, enabling a simultaneous excavation of wells at more than one locations, ahead of the tunnel face.

Thus, the wells could be dug in advance by independent teams and the course of the tunnel would be

defined by the array of the wells on the ground; the tunnel, which would have to meet the well bottoms,

was excavated step by step from one well to the next one uphill.

The basic components of the aqueduct, i.e. the tunnel and the wells, require craftsmen of different

qualities. The tunnel was probably excavated by one person at a time, assisted by a worker for removing

the rock cuttings. It is suggested that the wells were undertaken by a team of specialized craftsmen, as in

the Laurion mines.36

An intriguing similarity with the Hadrianic aqueduct can be recognized at the ancient Raschpëtzer

aqueduct in Luxemburg which was also designed within the aquifer, which is a sandstone layer interbedded

within marls. The total length is 600m and the wells reach the depth of 36m, contrary to other aqueducts

of the region which are shorter in length, with wells only 6 -10m deep. The sandstone aquifer yields 180m3

daily. The aqueduct is dated via dendrochronology in AD 130, i.e. during the Hadrianic period (however

with a rather wide age error margin).

Following the false trend of considering any shafts-and-gallery aqueduct as qanat, the Raschpëtzer

aqueduct is reported as such, although there is neither a single mother well nor an indication of Persian

influence and the climate is far from arid.37 I would suggest that it has similarities with the Roman

aqueduct at the Colony Caesarea Mauretaniae (modern Cherchel) in Algeria and the Hellenistic aqueduct

in Alexandria of Egypt. Indeed, Philip Leveau has underlined the importance of Roman hydraulic works in

Algeria, in the area of Cherchel, and the remarkable analogies with the Hellenistic galleries in sandstone

layers in Alexandria, a center of important innovations in hydraulic engineering.38 He also connects the

construction of the hydraulic installations in Cherchel with the migration of Greek engineers from Egypt.

Thanks to accurate underground surveying and geological mapping at the Raschpëtzer aqueduct, it

became clear that the tunnel was excavated stepwise, starting from the bottom of a well near the tunnel

entrance and heading upstream to the bottom of the next well. Errors in the heading of the tunnel

between two wells were adjusted by joining the tunnel section in progress to the next well, through

an S-like path to correct the error in the X-Y plan and via a ‘stairway’ to correct the error in the Z axis.

Similar S-like tunnel sections and gaps in elevation on both sides of a well are also reported by Defteraios,

regarding the northernmost section of the tunnel of the Hadrianic aqueduct.39

The S-shaped sections of the Hadrianic tunnel resulted similarly from the correction of the tunnel course

in order to meet the next well upstream. In parallel with tunneling, a separate team excavated the wells

ahead of the tunnel face, at locations and depths accurately determined by the surveyors’ team. In case

of tunnel deviation from the correct heading an S-shaped curved paths was necessary to correct for the

deviation and meet the next shaft. When the lateral deviation was small and the tunnel section passed

higher or lower than the shaft bottom, then the tunnel was enlarged vertically to pass at the scheduled

elevation; in this case, there was only an elevation difference on both sides of the wells. The larger section

of the tunnel occurred downstream when the tunnel crossed the shaft too high and vice versa.40

Hadrianic distribution network in Athens

Recent excavations at center of modern Athens, particularly in the National Garden (NG in Figure 6a)

and west of the Parliament, near the point H1, unearthed numerous remains of ancient hydraulic works.

Kakavoyiannis and Koursoumis 2013: 83. Since the end of the 6th century BC, craftsmen of the Phrearrians Demos of Laurion, Attica were

specialized in the construction of mining shafts by contact.

Kayser and Waringo 2003: 278.

Leveau 2012: 157. In the Roman colony Caesarea Mauretaniae, an area with an average annual precipitation of 633mm, the Roman galleries

capture water from a layer of sandstone and remained in use in modern times.

http://urbanspeleology.blogspot.gr/p/blog-page_2754.html

Defteraios et al. 2017: 52.

Great Waterworks in Roman Greece

This is not unexpected since the area lies at a higher altitude in relation to the ancient city, in the route

of ancient aqueducts; actually, all ancient aqueducts cross this area – except of the Acharnian aqueduct.

Although the course of the Hadrianic tunnel is accurately known, its continuation beyond the Lycabettus

reservoir up to the National Garden has not been identified yet. Ziller supported that the water was

conveyed from the reservoir to the city through lead pipes, one of which was found in situ (diameter

18cm, thickness 3cm).41 No more lead pipes have been reported and this single lead pipe could be simply

part of a fountain.

Branches of the Hadrianic aqueduct in Athens

Leigh described in detail the remains of the Hadrianic aqueduct in the Agora, by expertly combining the

ancient structures next to the City Eleusinion (H5, Figure 6a), with petrographic study of the mortars.42

The easternmost section consists of a channel covered by a vault, formed by 3 curved bricks held in place

by mortar. The channel of the aqueduct is built in opus mixtum and measures 0.60m wide in the interior

and 0.80m to the top of the side walls.43

The western extension of the channel is of different masonry and rests on a wall of poros blocks founded

on concrete bedding. Because of the steep slope of the ground, the northern continuation of the aqueduct

channel was carried on an arcade. The arcade was abandoned and replaced by a continuous foundation

wall that carried a new channel. The abandonment of the earlier channel was probably caused by the

Herulian sack of AD 267, and the reconstruction was probably connected with the revival of the Agora

area, around AD 400.

The continuation of the channel beyond the SW fountain at the Athenian Agora is unknown, but a similar

vaulted conduit with walls in opum mixtum seems to belong to the water distribution network44 (C6, Figure

6a). It was abandoned in the 4th century AD (as dated from the numerous lamps found inside and around,

from an adjacent workshop).

Another section of the Hadrianic aqueduct (H3, Figure 6a) was discovered in 1926, during excavations at

the former Royal Stables. It was first reported as a Roman aqueduct lined with clay plates and contained

many shafts.45 It is briefly described by Semni Karouzou as a probable part of the aqueduct constructed

by Hadrian.46 It was discovered south of the ancient city wall and was casually repaired in Late Antiquity.

The body of a Hellenistic statue and pieces of a 4th century AD sarcophagus were used to replace bricks

and plug a hole at the aqueduct.

Sections of the Hadrianic Aqueduct have been encountered in various excavations, but not recognized

as such yet. The conduit excavated by G. Mylonas west of the Parliament (1.7m below the solid rock and

close to a hypocaust), constitutes in my opinion part of a primary branch of the Hadrianic aqueduct

(H1).47 The construction of the conduit is similar to that of the inclined gallery at the northern end of the

Hadrianic tunnel, and the vaulted roof is composed of three curved bitched bricks. According to Mylonas

an extension of the conduit was also found in the center of the square in front of the National Monument

and at the NW end of the National Garden. Ventilation and cleaning shafts were also recorded at regular

intervals.

Ziller 1877: 120.

Leigh 2000: 118.

Picture of the vaulted channel at the Agora excavation and plan of the aqueduct respectively are displayed at: http://agora.ascsa.net/id/

  



Archaiologikon Deltion 

Archaiologikon Deltion 

Amandry 1947-1948: 386 fig.1, 391.

Mylonas 1932: 45.



Figure 6. Projection on Ktimatologio air photos: (a) Hadrianic aqueduct remains (H1-H6), above and (b) heavy

vaulted conduits of Roman concrete (C1- C8), below.

Great Waterworks in Roman Greece

Recent excavations confirm Mylonas’ description. A 27.3m long conduit built in opus latericium, along

which two shafts were interpolated, was excavated at the entrance F of the Syntagma Metro Station.48

The channel (H2, Figure 6a), is apparently the extension of the conduit described by Mylonas, directed

indeed towards the NW corner of the National Garden; it was however found open, without any roof.

The section H4 in the Figure 6a is hypothetical, drawn on the basis of geomorphological criteria along the

local crest line in combination with the distribution of Roman baths in the area.

Ziller described another conduit (H6, Figure 6a) which, I would suggest, could belong to a branch of the

Hadrianic aqueduct.49 He reported that during excavations searching for hydraulic works, conducted by

the Director of the Royal Garden (and ordered by Queen Amalia), a conduit was discovered behind the

scene of the Dionysius Theater, which was constructed in the system 12 and 18 depicted in Ziller’s table

VIII.50 Ziller described the characteristic roof of three curved bricks, but he assumed that the conduit

belonged to the aqueduct of the Royal Garden and that it was refurbished, when the theater was extended.

Secondary distribution network

The description of pipelines or conduits in excavation reports is in general either circumstantial or

omitted. Furthermore, some confusion often occurs regarding their function -as in the case of certain

heavy, vaulted conduits made of Roman concrete. A typical example is the vaulted conduit C1 (Figure 6b),

c. 72.5m long (dipping from North to South), which was excavated in two periods at the higher NE area of

the National Garden, in the yard of the Presidential Garrison.

As shown from the first excavation period, the conduit was founded in a trench on Athenian schists, 0.50m

below the surface. This conduit, had walls 0.55 -0.60m thick and external width 2.05m; the excavators

ascribed it to the Hadrianic aqueduct.51

Another section of the same vaulted conduit C1 was unearthed during the second excavation period;52

it was interpreted as a sewer belonging to a wide network located at many places in ancient Athens,

and was dated in the early Roman period. The argument was based firstly on the fact that six terracotta

pipes emptied into the conduit and secondly on pottery findings dated in the 3rd or early 4th century AD,

indicating the period of the conduit abandonment.

In my opinion, drainage from an area outside of the city walls (Figure 6a), would be more convenient in the

opposite direction, outwards, along a path following the modern Queen Sophia’s Avenue. In addition, a

terracotta pipeline composed of U-shaped elements would be sufficient for sewerage instead of the heavy

concrete conduit.53 Furthermore, in a drainage network the size of the ducts increases progressively;

in this specific case the conduit is oversized from the beginning and too heavy to belong to a sewer.

Therefore, I would suggest, that this conduit should be associated with a secondary network of water

distribution of the Hadrianic aqueduct.

A section of a similar vaulted conduit, C3, was excavated 0.30-0.55m below the surface (Figure 6b); it

is made of stonewalls bound with hydraulic mortar and measures 4.40m in length, 2.10 in width and

2.15 -2.25m in height. It is attributed by the excavator to a network of the Imperial period (2nd-3rd

century AD). 54 The numerous remains of conduits, similar in construction and contemporary, support

the assumption that they belong to a common network. The relation of this particular conduit to a water

supply network is indicated by an adjacent water reservoir (14.60 x 6m), excavated by the P. Bouyia

Archaiologikon Deltion 

Ziller 1877: 115.

Also, Dörpfeld and Reisch 1896: Table I.

Archaiologikon Deltion 

Archaiologikon Deltion 

See also Chiotis and Chioti 2014: 316.

Archaiologikon Deltion 



downwards of the previous discussed location.55 The reservoir, whose internal sides were coated with a

hydraulic plaster, was abandoned in the 4th century AD.

A similar vaulted conduit, C5, with plaster coated walls, dated in the 2nd to the 3rd century AD and of a

total known length of 30m, was unearthed close to the surface in three adjacent excavations.56 The floor

is covered by a layer of stones bound with plaster.

All the above massive and vaulted conduits share certain common characteristics: they have similar

massive dimensions; their walls are c. 0.50m thick, made of Roman concrete; their internal surface

is plastered and are dated mainly in the 2nd century AD. In addition, they are built near the surface,

apparently in a trench founded on the Athenian schists and were abandoned in the 4th century AD.

Therefore, we may conclude that they belong to a secondary network of water distribution fed by primary

branches of the Hadrianic aqueduct.

Having established the typical features of this network, some more conduits can be attributed to it: the

47m long conduit C2, constructed close to the primary branch H2;57 the 20m long conduit C4, built close

to a significant water reservoir,58 built on the fill of the moat (trench) of the classical precinct wall after

Sulla’s sack in 86 BC and on top of the Vallerian Wall constructed a few years prior to the Herulian raid in

AD 267; the previously mentioned vaulted conduit C6.59

In addition, the vaulted conduit C8 (2nd century AD), excavated along the Demosion Sema, presents

some common features.60 It was built at a depth of 3.40m at the same location with a 5.60m long, late

Hellenistic pipeline. This pipeline was composed of symmetrically coupled U-shaped terracotta parts

at a depth of 4.90m, well protected below two layers of amphorae; it was refurbished and reinforced by

curved terracotta parts commonly used for lining water wells. The pipeline was equipped with an access

shaft 0.90m in diameter, and was considered as a sewer by the excavator. I believe that the pipeline of

U-tubes and the access shaft were carefully designed as a branch of an older aqueduct directed towards

the Academy, similar in design to the Acharnian aqueduct. It was most likely built to replace the Kimonian

branch of the Peisistratean aqueduct mentioned earlier. The Acharnian type aqueduct was later replaced

by the vaulted conduit of Roman concrete, with 0.85m thick walls and 1.50m internal height, which was

again abandoned and plugged in the 4th century AD.

Another vaulted conduit, C7, which can possibly be attributed to the same network, is a heavy construction

of monumental dimensions, found near a Roman hypocaust.61

The co-existence of hydraulic lines of various dates is a common observation at many locations and

indicates that generally the courses of water supplying systems developed along the same routes

throughout the centuries. The distribution of water to a network of lower level was probably regulated

by reservoirs, as suggested by the excavation of some cisterns adjacent to the network. This holds at least

for the reservoir in the NE corner of the Syntagma square, found 2.5m below the surface, and coated

internally with a hydraulic plaster containing ceramic powder.62

Apart from the water distribution, another function of the secondary network would be managing the

water overflow of the Hadrianic aqueduct. The study of the ‘First Maps of the City of Athens’ by Prof.

Manolis Korres, that includes both ancient waterworks and the network of fountains of Athens in the

Archaiologikon Deltion 

Archaiologikon Deltion 

B’1, 2000: 78-80 note 27, figs. 10-11.

Archaiologikon Deltion 

Archaeologike Ephemeris

Archaiologikon Deltion 

Archaiologikon Deltion 

Archaiologikon Deltion

Archaeologike Ephemeris

by P. Bouyia (see above, note 55: Archaiologikon Deltion

Great Waterworks in Roman Greece

Ottoman period, has been a great contribution in many aspects. Korres was in position to locate on the

maps the sites of two important Roman cisterns both known under the common name Chamosterna.

They are distinguished in this study by the names Chamosterna 1 and Chamosterna 2. According to

Kordellas, the Chamosterna 1 at Kolokynthou was huge;63 it measured 22.5x28.0x3.0m and was similar

in construction with the arcade water bridges at Philothei and Perissos. Because of that he assumed

that the ancient aqueducts passing over these water bridges supplied water to the Chamosterna 1, at

Kolokynthou. The second cistern, Chamosterna 2, was fed by the Ilissos’ aqueduct (Figure 7) – also known

as aqueduct ‘tou Vounou’- which is described by Ziller and depicted on the Curtius and Kaupert map of

Athens.64

I consider the above cisterns to be in proper locations to collect the redundant water overflowing from

the vaulted conduits of the described secondary network (Figure 7). In particular, the conduits near the

Ilissos riverbed could feed the Chamosterna 2; the later could at the same time collect water from the

Ilissos’ aqueduct possibly transferring it to Piraeus.

As Leigh pointed out, according to Vitruvius, conduits brought water from Athens to Piraeus; however

this water was not drinkable: ‘No one drinks from it, but they use it for baths and so forth. They drink

from wells and thus avoid its ill effect’. 65

In addition to the Chamosterna cisterns shown in Figure 7,66 another couple of ancient cisterns, Mourouzi,

at the western extension of the Mourouzi Street, and NG, in the National Garden, are also marked. The

Kordellas 1879: 91.

Ziller 1877: 108; Curtius and Kaupert 1881; Korres 2010: 99.

Leigh 2001: 65.

Curtius and Kaupert 1881.

Figure 7. Ancient cisterns Chamosterna 1 (Cham1), Chamosterna 2 (Cham2), Mourouzi (M) and National Garden (N); vaulted conduits

C1 to C8; the Hadrianic aqueduct; the stone-built channel from the Haghios Demetrios ‘spring’; the Tzakoumakou channel; the

Ilissos’ aqueduct (‘tou Vounou’); the distribution channels I, II, and III supplying fountains during the Ottoman period.



Mourouzi cistern is depicted on the Kleanthis-Schaubert 1:2000 map, according to which Korres estimated

its size at 8.0 x 5.5m.67

According to F. Stauffert, the water supply system of Athens in the 1830s was fed by two aqueducts;68

namely the stone built channel from Haghios Demetrios and the aqueduct of Tzakoumakou, which are

marked upon the ‘First Maps of Athens’ (Figure 7). The water was collected in the Mourouzi cistern near

the Royal Palace and was distributed in Athens through 3 branches, also marked in Figure 7 (I, II and III).

The Mourouzi cistern is also depicted on the Sommer’s map of 1841, along with the supplying aqueduct

directed from east to west.69

Based on the above, it is suggested that even prior to the cleaning of the Hadrianic aqueduct in the 19th

century, the water supply Athens in the Ottoman period was based on ancient hydraulic works.

It is inferred that the secondary network of vaulted conduits was abandoned in the 4 century AD. This

does not necessarily imply that all the primary branches of the Hadrianic aqueduct ceased supplying

water as well. For example the primary branch supplying the Agora might be longer in operation. The

construction of the late Roman aqueduct, which supplied the water mill in the Agora, strongly indicates

abandonment of the Hadrianic aqueduct in the 5th century AD. Coins, pottery, and lamps found in the

mill provide a date for the mill construction at the third quarter of the 5th century AD.70

Tertiary network of water distribution

The tertiary distribution network, supplied by the secondary one via cisterns, is generally composed of

terracotta pipelines, the commonest of which are made of coupled U-shaped parts, forming an elliptical

duct. Actually, similar ducts are used both for water distribution and for waste water removal. A typical

example of tertiary water distribution is the combination of various duct types excavated along an

71

The Roman baths below the church of Haghios Nikodemos were apparently supplied by two alternative

systems. There are actually two baths built at different levels and perhaps at different periods. The

northern one, which is better preserved and dated by Travlos probably in the mid-second century AD, was

supplied by an elliptical pipeline, presumably fed by the Hadrianic aqueduct. 72 The second alternative,

associated with both baths is a well in possible connection with the underground National Garden-

Hymettos aqueduct.

It is suggested that traditional wells remained in use, particularly for drinking water, concurrently with

the Hadrianic aqueduct.

Tradition of hydraulic technology in Greece before the Hadrianic aqueduct

As perceptively underlined by Leigh, in regard with the ancient hydraulic technology, it is necessary to

think in terms of a continuous engineering tradition, extending from Assyria through the works of the

Nabataeans, to the Greeks, the Carthaginians, and the Etruscans, then to the Romans and beyond.73 It is

therefore interesting to track the possible contribution of the Greek tradition, already well developed

since the Mycenaean period, to the design of the Hadrianic aqueduct and hydraulics engineering in

general.

Korres 2010: 99.

Stauffert, as cited in Papageorgiou-Venetas 2010: 109.

Sommer 1841, as accessed from http://www.nhmuseum.gr/el/fakelos-syllogon/antikeimena/16042_el/.

Franz et al. 1988: 81.

Archaeologike Ephemeris

Archaeologike Ephemeris

Leigh 1998: 118.

Great Waterworks in Roman Greece

The military engineers of the Mycenaean period confronted the demand to supply fortified cities with

potable water, which could not be cut off or contaminated by the enemy during sieges. The Klepshydra

fountain in the Athenian Acropolis, the secret underground cistern at Mycenae and the underground

spring at Tiryns are representative examples of the engineers’ ingenuity for the water supply of a citadel.

Numerous remains of Mycenaean hydraulic works have been unearthed in the Boeotian Thebes, an area

with abundant springs in antiquity. An interesting hydraulic structure was excavated by Keramopoullos

in the citadel in the so called ‘Kadmos House’, the most important archaeological site at Thebes.74 A well

was found near a Mycenaean pottery kiln; at its bottom (at a depth of 3.60m), a rock-cut duct 1.10m

high and 0.50m wide was ramified in 5 directions. This structure probably constituted a groundwater

collection work, a technique applied later in the Classical period, e.g. in the Corinthian wells.75

Early tunneling techniques in Greece were also partially affected by the underground burial structures

(and vice versa). The similarity of the hydraulic galleries at Thebes to the ‘dromos’ of the Mycenaean

chamber tombs has already been emphasized by the excavators.76 In addition, peculiar chamber tombs

on Aegina Island confirm a long tradition in underground works since the 8th century BC; they were

accessed by shafts and in some cases were more complicated in structure than the shafts-and-gallery

technique.77 The geological conditions were favorable on Aegina, where solid layers of limestone (the

roof of the tombs) covered soft layers of marls, within which the chambers were easily carved.

At Corinth, underground waterworks capturing spring aquifers were cleverly adjusted to the local geology

and were applied from the early 6th century BC onwards. Vertical shafts were dug at regular intervals to

facilitate rocks removal and improve ventilation. The shafts had toe-holes for climbing and their long

dimension was aligned with the tunnel’s direction,78 a pattern adopted also in the Hadrianic aqueduct.

The shafts-and-gallery tunneling technique has been applied rather early in Greece. The Eupalenian

aqueduct on Samos Island is dated in 550 BC.79 It is composed of four differently constructed sectors. The

first one is a stone built surface channel carrying spring water to an underground gallery (built in the

shafts-and-gallery technique) which in turn transfers the water to the famous Eupalinos tunnel. The final

sector conveys the water to the Pythagoreion harbor through a shafts-and-gallery tunnel, c. 500m long.80

It should be noted here Graeber’s view that the Aegina and Megara aqueducts were the models for the

Eupalenian aqueduct.81

Two aqueducts of the shafts-and-gallery type, dated in the period of Thebe’s prosperity in the Classical

period were also excavated.82

The National Garden-Hymettus aqueduct is another underground aqueduct tapping groundwater. It was

constructed following the shafts-and-gallery tunneling technique, perhaps in the 4th century BC and,

despite the absence of any springs, continues to irrigate the National Garden of Athens.83

The Acharnian aqueduct of the 4th century BC constitutes, in a sense, an improvement of the Archaic

Peisistratean aqueduct; the latter was built in Athens in the late 6th century BC and conveyed spring

water through a terracotta pipeline with openings, placed in a cut and covered trench.84 In the Acharnian

aqueduct mortared bricks along the straight joint of the U-shaped drums and plastering of the elliptical

junction can compose a watertight terracotta pipeline. Such a structure was installed in the branch D

Keramopoullos 1911: 149.

Landon 2003: 44.

Papadakis 1911: 132-142.

Papastavrou 2006: 129.

Hill 1964: 54.

Kienast 1995: 186.

Zambas et al. 2017: 64.

Graeber 1905: 558.

Symeonoglou 1985.

Chiotis and Chioti 2012: 408; Chiotis and Marinos 2012: 23.

Chiotis and Chioti 2012: 424.



of the Megara aqueduct, to avoid water loss in the gallery and keep the water clean.85 The waterline of

the Late Roman aqueduct in Athens, to which the Perissos and Philothei water bridges belong, was also

similar.

Proper siphons are not known in Attica, but they are known in Crete – in the Hellenistic aqueduct of

Lyttos and the Roman aqueduct of Gortyn.86

Inclined wells along the aqueduct gallery were introduced to the Hellenistic aqueduct on Rhodos Island.87

A unique and ingenious combination of an inclined gallery with three vertical wells for water uplift is

well preserved at Perachora in Corinthia.88

The different attitude and scale of hydraulic works of Greeks and Romans is best exemplified at ancient

Dekapolis. Until the Hellenistic period water supply of Gadara was

based on cisterns and a few springs. Economic prosperity and population increase during the late 3rd

century BC necessitated the construction of a shafts-and-gallery aqueduct, 30km long with 700 shallow

wells, for the transportation (in a terracotta pipeline) of water from a spring, yielding about 4l/sec. To

pass the tunnel through a ridge up to 72m below the surface, over a distance of 600m, an auxiliary gallery

was built, a few meters apart, connected with the aqueduct gallery with short inclined shafts dipping

about 45o.89 As I consider, this was an ingenious solution for creating a ventilation circuit at the gallery

faces, an approach resembling mining technology.

   

and 2nd centuries AD

with about 2900 inclined shafts. Undoubtedly, it constitutes the most complex, ambitious and impressive

ancient aqueduct. It is noted that the inclined shafts were dug out initially from 20 to 200m intervals and

at a depth from 5 to 70m. Then, from the shaft bottom, tunneling started with pilot tunnels in opposite

directions, which were enlarged to the final dimension after the meeting of the tunnel sections, a step

which permitted the correction of any deviations from the planned geometry.

Based on the above examples, one can discern that groundwater hydrogeology and tunneling in the

shafts-and-gallery technique were mastered in Greece in a much earlier period, already in the 6th century

BC. However, there is another common parameter for the completion of the above discussed aqueducts of

such complex geometry (i.e. the Eupalenian tunnel, the underground section of the Hellenistic aqueduct

of Gadara, the Hadrianic aqueduct Athens and the Roman aqueduct of Dekapolis), and that is the

application of mathematics: as already emphasized by Theodosios Tassios, building the Eupalinos tunnel

would not be possible without applying the Pythagorian mathematics. Similarly, the route of the Roman

Dekapolis aqueduct was fixed using triangulation, as known from Pythagoras, Hipparchos of Nicaea,

Menelaus of Alexandria and Heron.90 In both the Eupalinos and the Decapolis aqueduct, measurements

were written on the rock wall in Greek letters – an indication of the Greek origin of their surveyors.

Summarizing the above discussion, the Hadrianic aqueduct in Athens is a hydraulic work built in the

Roman period, but technically based on the long hydraulic tradition in Greece and Asia Minor – in

combination with Hellenistic mathematics and the mining technology of deep shafts (Laurion). A transfer

of hydraulic models from the Middle East, or elsewhere, cannot be excluded in the 7th or early in the 6th

century BC, especially in Aegina and Megara, but cannot be traced. From the 6th century BC onwards, up

to the Roman revolution in civil engineering works (e.g. the introduction of Roman concrete, exemplified

by the impressive water bridges), hydraulic engineering developments in Greece must be considered

indigenous.

Avgerinou 2017: 47.

Lyttos: Oikonomakis 1984: 66-99. Also, Amanda Kelly’s paper (in present volume). Gortyn: Giorgi 2007: 287–320.

Voudouris et al. 2013.

Tomlinson 1969.

Döring 2017: 178.

Döring 2017: 189.

Great Waterworks in Roman Greece

Concluding remarks

Water supply of modern Athens was based on ancient aqueducts until the Marathon reservoir was

constructed in 1935. Even before the repair of the Hadrianic aqueduct in the 19th century, the small

city of Athens benefited for centuries from the ruined ancient aqueducts. The impressive network of

fountains of the Ottoman period – recently revealed thanks to the meticulous study of the ‘First Maps of

Athens’ by Korres91 – was fed by overflows both from the Hadrianic aqueduct at the Haghios Demetrios

Church (Figures 1 and 7) and from the National Garden-Hymettus aqueduct, a section of which is known

near the Haghios Thomas Church, at Goudi. In addition, the National Garden continues to be irrigated

from the National Garden-Hymettus aqueduct.

During the Ottoman period the stone-built channel from Haghios Demetrios followed a smooth course

towards Athens (Figure 7), thus indicating that the construction of the reservoir on the Lycabettus Hill

was not a necessity; it was however in accordance with the Roman concept of castellum aquae, located in

Athens at an imposing location, selected also for the demonstration of its monumental façade.

The tunneling techniques made significant progress in Greece in areas rich in springs, such as Thebes

and Corinth, where the shafts-and-gallery aqueducts gradually developed as an extension of the simpler

previous techniques (that of a single well with galleries at the bottom, and spring-tunneling).92 The

gradual development of a short spring-tunnel into a complicated network of tunnels and shafts in only

a couple of centuries can be better observed at the Peirene fountain in Corinth. The tunnels followed

the sinuous aquifer in a mining-like technique, without a predefined plan. The excavation of shafts and

galleries was adjusted to the local geological conditions and the winding course of the aquifer.

The sustainability of the ancient aqueducts in Athens until today is due to their adequately deep gallery

and shafts which enter into the water table. The archaic aqueducts at Aegina and Megara (6th century

BC), were the earliest of this type in Greece, even earlier than the outstanding Eupalinos aqueduct (c. 550

BC), the end sectors of which were also constructed in the shaft-and-gallery technique.

An important particularity of the Hadrianic aqueduct is the significantly greater depth of the shafts,

up to c. 41m. Over significant part of its course, the tunnel and the wells were dug below the water

table (for improved groundwater capturing). This required an innovation in tunneling for the drainage

of the groundwater from the excavation front. The Hadrianic aqueduct was actually a development of the

shafts-and-gallery technique (as used in the Aegina, Megara and National Garden- Hymettus aqueducts),

applied to greater depths.

Neither the above aqueducts, nor the Hadrianic one, can be considered as qanats (cf. the definition and

diffusion of which is examined in the Appendix). One reason is because the shafts-and-gallery technique

was applied in Greece earlier than the Achaemenid Empire (538/532-332 BC) – which is credited with

the development and diffusion of the Iranian qanats. Secondly, the Greek shafts-and-gallery aqueducts

capture groundwater almost along their entire course, whereas Iranian qanats are typically fed from a

mother well and their transportation course crosses dry rocks in arid areas. And, lastly, because aqueducts

such as, the Eupalenian in Samos, the Hellenistic at Gadara, the Hadrianic at Athens and the Roman

Dekapolis aqueduct were built according to the advanced applied mathematics of their time. By contrast,

the surveying of the Iranian qanats is a simple empirical exercise.

In conclusion, as demonstrated in the examination of the hydraulic tradition in Greece, the Hadrianic

aqueduct is the evolutionary product of mathematics and indigenous hydraulic and mining technology.

It could be argued that the prefabricated terracotta pieces for roof support constitute a Roman influence,

not without a counterpoint, but this is an issue of minor importance.

Korres 2010: 93.

The final constructions varied in form due to the different local conditions: straight aqueducts with sometimes overlapping tunnels at Thebes

and meandering network of galleries at Corinth.



There are indications that similarly to the Roman Raschpëtzer aqueduct in Luxemburg, the Hadrianic

tunnel was excavated at intervals upstream, starting from a well to the next one uphill. Furthermore, it is

inferred that the construction of the Hadrianic aqueduct was undertaken simultaneously in more sectors

starting from lateral deviations from the tunnel at the surface and by digging upstream. Thus, the water

was drained by gravity flow and the water table was depressed and the excavation of more adjacent wells

below the water table was possible. The wells were sunk in advance defining the course of the aqueduct

and the tunneling followed at close distance.

The water distribution network from the Hadrianic aqueduct to ancient Athens can be traced from patchy

– but structurally consistent - remains. The primary distribution netwok consisted of shallow tunnels,

a couple of meters below the surface, exemplified in construction details by the Agora branch, near the

City Eleusinion. Two main branches can be traced, one passing west of the parliament along the Amalia’s

Avenue (H1, H2, H4 in Figure 6a), which was subdivided in two sections at least, one directed to the Agora

(H5) and the other one SW of the Dionysus Theater (H6). The second branch of the primary network was

traced only at one location (H3), but it is guessed that it was directed towards the Plato’s Academy (based

on the remains of the secondary distribution netwok).

The latter one is represented by remains of vaulted tunnels of Roman concrete, with thick walls,

constructed in shallow covered trenches. The secondary distribution netwok was constructed in the 2nd

century AD and abandoned in the 4th century. It also served for collecting the water overflow to large

cisterns, the largest of which was the Chamosterna at Kolokynthou, in the middle of a fertile land. The

second cistern, near the Ilissos River, could collect the redundant water of the Hadrianic aqueduct in

this region and the water from the Ilissos’ aqueduct and transfer it, possibly, to Piraeus. The tertiary

distribution netwok consisted mainly of terracotta pipelines (or built channels), fed from cisterns. The

co-existence of hydraulic lines of various periods indicates that the courses of the main hydraulic works

were developed along the same routes throughout the centuries. The operation of a new Late Roman

aqueduct in the 3rd quarter of the 5th century AD for the supply of the Agora water mill indicates the

abandonment of the Hadrianic aqueduct before that.

Acknowledgements

Professor Theodosios Tassios, President of the Association of Ancient Greek Technology Studies (EMAET),

who has promoted for many decades the study of ancient Greek technology, has constantly inspired

me to study the ancient hydraulics. He also contributed through his expert advice on the improvement

of this study. The Chief Editor of this volume, Dr Georgia Aristodemou, improved my original text and

kindly tolerated my corrections. My collaboration with Panayiotis Defteraios contributed significantly to

my understanding of the underground structure of the Hadrianic aqueduct, thanks to his speleological

investigations and his kind cooperation. Effie Nestorides (EYDAP), third generation Engineer, of a family

with great contribution to the study and promotion of the Hadrianic aqueduct, has been constantly

supportive and encouraging for many years.

My long involvement with the study of the ancient hydraulic works would not be possible without my

family’s appreciative and unceasing encouragement.

Addressing my thanks to all of them is a great pleasure.

Great Waterworks in Roman Greece

Appendix

Comments on the nature and definition of qanats and comparison to the Hadrianic aqueduct

Following Antisthenes, the definition of the various types of underground aqueducts is here investigated,

with emphasis on qanats. The distinction between qanats and underground aqueducts is often

misunderstood and because of that there is an increasing tendency of naming qanat any shafts-and-

gallery aqueduct. Thus, qanats are incorrectly defined as ‘gently sloping, artificially built subsurface

channels, which lead groundwater from the margins of the mountains to the area where water is needed,

sometimes many kilometers away’.93 Obviously, this broad definition includes practically almost all the

underground aqueducts.

The same definition was recently applied to the ancient Greek underground aqueducts, which were

characterized as aqueduct-like qanats – a neologism that intensified the confusion. Strangely enough, it

was further argued that ‘when comparing the etymology/definition of the terms aqueduct and qanat it

should be effortlessly concluded that Qanat is identical to Aqueduct, or at least it is included therein’.94

May combined technical description and rather doubtful historical assumptions to define the qanat

95 Accordingly, a

qanat consists of an underground tunnel which uses gravity to convey water from the water table (or

springs) at higher elevations, to the surface of lower lands. Qanats also have a series of vertical shafts

used for tunnel excavation, maintenance, ventilation and lighting. The oldest qanats have been found in

the northern part of Iran and date back to c. 3000 years ago, when the Arians (Aryans) settled in present

day Iran. From 550-331 BC Persian rule extended from the Indus to the Nile, during which time qanat

technology spread. 96

An essential and distinguishing technical feature of qanats omitted in this definition-but stated by

Lightfoot-is that ‘a qanat is a subterranean aqueduct engineered to collect groundwater from a mother

well’.97 To emphasize the importance of this comment, it is noted that during the 20th century the ‘mother

wells’ of the Iranian qanats were up to 150 m deep (in Tehran); however, the deepest ‘mother wells’ in Iran

could reach depths of 300 or even 320m.98

It is self-evident that qanats were expensive in construction and maintenance, and it might take 20 to 30

years to dig a new qanat.99 It is also significant that the term qanat is used in Iran in a particular cultural,

technical and geographical context. It seems, therefore, that ‘it cannot be used to designate all the types

of water capturing tunnels, in various periods, in diverse geographical and climatic zones’.100

Boucharlat proposed the introduction of a more general term when describing tunnels capturing

underground water, which would also include the qanats. He preferred in French the term ‘galerie de

captage’.101

In this paper, following Wilson, the general term shafts-and-gallery aqueduct is used when describing

all hydraulic structures where an inclined tunnel along a ‘chain’ of wells uses the gravity based water

flow.102 This definition applies to all underground aqueducts irrespectively of the origin of water source.

It includes Assyrian tunnels transferring water from a river, the Hadrianic aqueduct in Athens tapping

groundwater almost all along its course, the falaj in the United Arab Emirates and Oman tapping shallow

Weingartner 2007: A1556.

Voudouris et al. 2013: 1341

Mays 2010: 3.

The period of Achaemenid Empire is defined from 538 to 332 BC by Magee (2007: 87) or from 532 to 332 BCet al. (2012: 167).

Lightfoot 2000: 215.

Wilkinson 2012:16.

Esfandiari 2007: 60.

Salesse 2001: 713.

Boucharlat 2000: 158.

Wilson 2008: 290.



aquifers along valleys of seasonal streams and the typical Iranian qanats which tap water from strongly

dipping aquifers in alluvial fans.103 In this way, the confusion of naming qanat any shafts-and-gallery

structure is avoided. Further subdivision of the shafts-and-gallery aqueducts should be based on the

origin of water, the depth and dip of aquifer, and historical, geographical and climatic data.

According to Wilson, the shafts-and-gallery tunneling technique -that many authors associate with

the Persian qanats – was firstly developed in the Assyrian Empire for the construction of long-distance

artificial canals supplying water usually from distant rivers. They crossed intervening ridges in the form

of tunnels several hundred meters (or even several kilometers) long. The shafts-and-gallery tunneling

technique solved the problem of underground surveying; the shafts were set at sufficiently close intervals

and thus, short, manageable sections were created. The Assyrian king Assurnasipal II (884-859 BC) dug a

canal 19.5km long from the Upper Zab River to supply Nimrud; it passed under a rock ridge at Negoub via

a tunnel 7km long, dug between pairs of vertical shafts. The system supplied the city of Nimrud and also

irrigated the fields around it.104

Physical geography and climatic change at the beginning of the 1st millennium BC are indispensable

factors both for understanding the causes that led to the development of falaj in the Peninsula of Oman,

and for defining falaj.

Boucharlat demonstrated that the groundwater draining galleries named falaj firstly appeared in the

present-day United Arab Emirates and Oman early in the 1st millennium BC; he associated the introduction

of the draining galleries with a dramatic growth in population.105 He also investigated the type of aquifers

and concluded that galleries in the United Arab Emirates were shallow, and tapped water from valley

aquifers or dry wadi. In contrast, the Iranian qanats appeared in the Partho-Sasanian period or even later

and collected water from deeper aquifers.

There was a rapid settlement during the Iron Age II (c.1000-600 BC) throughout the southeastern

Arabia. The falaj irrigation system is in close association with a number of piedmont settlements in the

southeastern Arabia dated in Iron Age II (based on calibrated radiocarbon dates reported by Magee).

The falaj irrigation permitted intensive and year-round agriculture in the small 20-30km strip of alluvial

piedmont that flanks the Hajar Mountains to the East and West. This dating also confirms that the falaj

system was not an Achaemenid-period technological adaptation from Iran.106

More than a dozen underground shafts-and-gallery aqueducts have been dated to the Iron Age or have

been directly associated with Iron Age settlements. Where the water source could be determined, it was

associated with wadi flow or superficial aquifers in small depressions receiving local runoff. It is also

noted that Iron Age galleries differ from later Islamic qanats, which are deeper galleries catching water

from deep aquifers.107

The shallow depth of the aquifer was clearly confirmed at the excavation of the Al Madam falaj system

(a water catchment gallery). The gallery was located 1.3m below the surface and originally 1.5m high;

subsequently the floor was lowered by 4.80m in order to reach the declining water table in a period of

drought, due to which both the falaj and the adjacent village were finally abandoned.108

The generally accepted view that the qanat was invented in the well-watered regions of northwestern

Iran around the 8th century BC and then diffused across the Middle East by the Achaemenid Empire was

In strongly dipping aquifers, as quite often happens in alluvial fans in Iran in the piedmont, along normal fault lines of significant vertical

displacement, the possible zone of water capturing with a slightly inclined gallery is very narrow and this explains why in the typical Iranian

qanats the aquifer is tapped by a single mother well only. In addition, deeper aquifers are most likely artesian. Iranian qanats combining these

features are distinguished here as typical ones.

Wilson 2008: 290.

Boucharlat 2003: 169.

Magee 2007: 90.

Mouton et al. 2011: 16.

Cerro 2012: 134.

Great Waterworks in Roman Greece

recently challenged via archaeological evidence. It is presently considered that the emergence of the

qanat is more likely to have occurred in the southern regions of Iran, Pakistan and Arabia.109 Palaeoclimatic

evidence suggests that at the beginning of the 1st millennium BC these regions suffered from declining

precipitation; this pattern does not apply in southeastern Arabia (United Arab Emirates and Sultanate

of Oman). Extensive fieldwork recorded a rapid and intense establishment of new settlements around

1000 BC. Archaeological research at several of these settlements revealed the existence of ‘qanat’ systems

dated at the first few centuries of the 1st millennium BC.110 These ‘qanat’ systems constitute the earliest

evidence of this irrigation system until today.111

However, the identification of falaj with qanat is a precarious simplification (given that falaj appeared

earlier than qanats and taps shallow open aquifers). Furthermore, three types of falaj are distinguished

in Oman, in relation to the water sources: ghaili falaj taps water from the gravel of a river or wadi; aini falaj

draws water from one or more natural springs and daudi falaj taps an underground aquifer.112

It is further emphasized that the conventional falaj in Oman (either ghaili or aini) constitutes the response

of small communities to the changing climatic conditions, by which surface channels were gradually

transformed into shallow underground galleries. By contrast, it seems that the typical Iranian qanat was

planned in advance and executed as a major project for tapping deeper water resources from confined

aquifers via a single mother well and this is their distinguishing feature.

Table 1 summarizes the basic features of the shafts-and-gallery aqueducts discussed previously. Particular

emphasis is given on the origin of water in order to indicate the significant existing variability. Among

the aqueducts listed, only the Xinjiang aqueduct in China can be considered as qanat of the Iranian type.

The impressive variability in both the period of first installation and the diversified water capturing

methods, results from the technological adjustment to local historical and environmental conditions

over a long period starting from the Bronze and the Iron Age. As expressed by Wilkinson,113 water-supply

systems are sensitively engineered to suit local hydrogeological conditions.

It is therefore a misrepresentation of historical evidence to merge all these types of aqueducts under

the common model of Iranian qanats

qanat can be directly dated from the pre-Islamic period and even from the following centuries, despite

some assumptions put forward from time to time, especially for the Khorasan province, corresponding

114

and-gallery aqueducts was very likely polycentric during varied periods of the 1st millennium BC. Much

later, the second generation might have been actually implemented in Iran around the middle of the 1st

millennium BC

Finally, Chiotis (2007) associates the development of early underground aqueducts with world-wide

     

the process of adaptation of communities to natural decline of precipitation and springs, particularly

intensive in the turn of the second to the first millennium BC. Thus, areas rich in water in the first half of

the Holocene were gradually transformed into deserts in Libya and Western Egypt, with few oases near

springs. Subsequently, the decline of the springs forced the invention of underground aqueducts, the

foggaras in Libya, where outcrops of fossil aquifers were the last refugia for settlements; in the Peninsula

of Oman, the Hajar Mountains were similarly the last refugia, thanks to their higher precipitation, which

Wilkinson et al. 2012.

The application by Wilkinson et al. (2012) of the term qanat instead of falaj is rather confusing, because of the existence of several different

types of falaj; Magee (2007), however, avoided this misconception.

Wilkinson et al. 2012: 168.

Al-Marshudi 2007: 34.

Wilkinson 2003: 45.

Boucharlat 2017: 296.



could support irrigation in narrow zones at the foothills of the mountain belt by means of underground

115

Table 1

Indicative types of the origin of water in shafts-and-gallery (SAG) aqueducts

Origin

of water

Typical SAG

aqueducts

Bibliographical

Reference Comments

Shallow unconfined

aquifers below dry

valleys (wadi)

Ghaili falaj at

Al Madam, UAE Cero 2012: 134

Iron Age II (c.1000-600

BC)

River

Assyrian tunnels Wilson 2008: 290 early 9th century BC

(first half)

Spring

1. Eupalenian tunnel (Samos, Greece)

Zambas et al. 2017

Horizontal blind tunnel

with SAG end sections,

550 BC

2. Qanāt Fir’aun, or Dekapolis aqueduct Döring 2017

Tunnel and inclined

shafts, 1st and 2nd

centuries AD

Confined aquifers 1.Peirene (Corinth, Greece)

Hill 1964 Twisting spring tunnels

and shafts, 6th c. BC

2. Raschpëtzer aqueduct (Luxemburg) Kayser and Waringo 2003 2. early 2nd century AD

(first half) ,SAG

3. Perachora aqueduct (Corinthia, Greece) Tomlinson 1969 Hellenistic. Inclined

gallery-vertical shafts

4. Rhodos (Greece) Voudouris et al. 2013

Hellenistic. Horizontal

tunnel-inclined shafts

Unconfined aquifer

(SAG at the level of

the water table)

Aegina, Greece

Megara, Greece

Graeber 1905: 558

Earlier than the

Eupalenian tunnel

Tunneling below

the water table

1. Hadrianic aqueduct of Athens

Chiotis, present paper

AD 125-140 shafts up to

41m

2. Hadrianic aqueduct of Athens, modern

branches 1 and 2 Chiotis, present paper 19th -20th century

Springs Aini falaj (UAE) UAE 2006: 5 Modern

Groundwater from

mother wells Daudi falaj (UAE) UAE 2006: 13

Modern, shafts 6.5 to

30m.

Mother wells in

steep confined

aquifers Qanat at Xinjiang (China) Hu et al. 2012: 216

Modern, shafts: 40-70m,

100m max

* Abbreviations used





Chiotis 2017: 4.

Great Waterworks in Roman Greece

Bibliography

Al-Marshudi, A.S. 2007. Institutional arrangement in the Sultanate of Oman. In: Proceedings of the

International History Seminar on Irrigation and Drainage: 31-42. Tehran: Iranian National Committee on

Irrigation and Drainage.

Amandry, P. 1947-1948. Chronique des Fouilles en 1946. Bulletin de Correspondance Hellénique

Angelakis, A. N., Mamassis, N., Dialynas, E. G. and Defteraios, P. 2014. Urban Water Supply, Wastewater,

and Stormwater Considerations in Ancient Hellas: Lessons Learned. Environment and Natural Resources

Research 4.3: 95-102.

  

100 χρόνια Αρχαιολογικού Έργου στη Θήβα. Οι πρωτεργάτες των ερευνών και οι

συνεχιστές τους15-17 Νοεμβρίου 2002, Αθήνα 2014: 1-57. Athens: Ministry of

Culture.

              

            

Αρχαιολογικές Συμβολές vol. A (Attica): 161-171. Athens: Museum of Cycladic Art.

Avgerinou, P. 2017. The Archaic underground aqueduct in Megara, in the Chapter 4: Updated Appraisal

of Ancient Underground Aqueducts in Greece. In: A. N. Angelakis, E. D. Chiotis, S. Eslamian and H.

Weingartner (eds) Underground Aqueducts Handbook: 44-48. Boca Raton: CRC Press.



questions sur leurs relations avec les galeries du plateau iranien. In: P. Briant (ed.) Irrigation et drainage

dans l’Antiquité, qanats et canalisations souterraines en Iran, en Égypte et en Grèce: 157-183. Paris: Thotm

Éditions.

Boucharlat, R. 2003. Iron Age water-draining galleries and the Iranian ‘Qanat’. In: D. T. Potts, H. Al

Naboodah and P. Hellyer (eds) Archaeology of the United Arab Emirates. Proceedings of the First International

Conference on the Archaeology of United Arab Emirates: 159-172. London: Trident Press Ltd.

             

Arab Emirates as Case Studies. In: A. N. Angelakis, E. D. Chiotis, S. Eslamian and H. Weingartner (eds)

Underground Aqueducts Handbook: 279-301. Boca Raton: CRC Press.

Cerro, C. del 2012. Some evidence of crisis and abandonment at the end of the Iron Age in Al Madam

1-Thuqaibah, Sharjah. In: D.T. Potts and P. Hellyer (eds) Fifty Years of Emirates Archaeology. Proceedings of

the Second International Conference on the Archaeology of the United Arab Emirates: 132-139., London/Dubai:

Motivate Publishing.

Chiotis, E. 2017. Underground Aqueducts: Types, Definitions and General Conclusions. In: A. N. Angelakis,

E. D. Chiotis, S. Eslamian and H. Weingartner (eds) Underground Aqueducts Handbook: 1-16. Boca Raton:

CRC Press.



HΑγοράστηΜεσόγειοαπό τους Ομηρικούς έως τους

Ρωμαϊκούς χρόνους, Διεθνές Επιστημονικό Συνέδριο//The Agora in the Mediterranean from Homeric

     

Culture.

Chiotis E. D. and Chioti L. E. 2012. Water supply of Athens in antiquity, In: A. N. Angelakis, L. W. Mays, D.

Koutsoyiannis and N. Mamassis (eds) Evolution of Water Supply Through the Millennia: 407-445. London:

IWA Publishing.

Chiotis, E. D. and Chioti, L. E. 2014. Drainage and Sewerage Systems at Ancient Athens, Greece. In: A. N.

Evolution of Sanitation and Wastewater Technologies through the Centuries:

313-329. London: IWA Publishing.

Chiotis, E. D. and Marinos, P. G. 2012. Geological aspects on the sustainability of ancient aqueducts of

Athens, Bulletin of the Geological Society of Greece, 46: 16-38. (http://www.geosociety.gr/images/news_

files/EGE_XLVI/2_CHIOTIS_p16-38.pdf).

Karten von Attika, Berlin: D. Reimer.

Defteraios, P., Chiotis, E. and Mamassis, N. 2017. Underground evidence on the tunneling technique of the

Hadrianic aqueduct of Athens, in the Chapter 4: Updated Appraisal of Ancient Underground Aqueducts

in Greece. In: A. N. Angelakis, E. D. Chiotis, S. Eslamian and H. Weingartner (eds) Underground Aqueducts

Handbook: 50-54. Boca Raton: CRC Press.



                

Angelakis, E. D. Chiotis, S. Eslamian and H. Weingartner (eds) Underground Aqueducts Handbook: 173-

196. Boca Raton: CRC Press.

Dörpfeld, W. and Reisch, E. 1896. Das Griechische Theater: Beiträge zur Geschichte des Dionysos-theaters in Athen

und anderer griechischer Theater. 

Esfandiari, O. 2007. Qanat: Iranian’s most remarkable ancient irrigation system. In Proceedings of the

International History Seminar on Irrigation and Drainage: 55-61 Tehran: Iranian National Committee on

Irrigation and Drainage.

Fahlbusch, H., 2014. The Roman Water Supply System of Pergamum. In: Ch. Ohlig and Ts. Tsuk (eds) Cura

Aquarum in Israel II: Water in Antiquity. Proceedings of the 15th International Conference on the History of Water

Management and Hydraulic Engineering in the Mediterranean Region Israel 14-20 October 2012 (Schriften der

Deutschen Wasserhistorischen Gesellschaft 21): 165-184. Siegburg.

The Athenian Agora, XXIV. Late Antiquity: AD 267-700: 1-156:

Princeton: American School of Classical Studies at Athens.



  Αρχαιολογικές Συμβολές 

Athens: Museum of Cycladic Art.

Giorgi, E. 2007. Water technology at Gortyn in the 4th -7th c. AD: transport, storage, distribution. In: L.

Lavan, E. Zanini, and A. Sarantis (eds) Technology in Transition A.D. 300–650: 287-320. Leiden: Brill.

Graeber, F. 1905. Die Wasserleitung des Peisistratos und die Wasserversorgung des alten Athen. Zentralblatt

der Bauverwaltung 25: 557-560.

Hill, B. H, 1964. Corinth, I. 6. The Springs: Peirene, Sacred Spring, Glauke. Princeton: American School of

Classical Studies at Athens.



characteristics and modern implications for environmental protection. Journal of Arid Land

           

Archaeologike Ephemeris 152: 77-102.

Karouzos, Ch., 1926. Thebes. Archaelogikon Deltion

Kayser, P. and Waringo G. 2003. Die unterirdische Wasserleitung der Raschpëtzer. Ein Monument antiker

Ingenieurbaukunst aus Luxemburg. In: Wasserversorgung aus Qanaten – Qanate als Vorbilder im Tunnelbau,

Internationales Frontinus-Symposium, Walferdange, Luxemburg (Schriftenreihe der Frontinus-Gesellschaft

vol. 26): 277-292.

           Praktikatēs en Athēnais

Archaiologikēs Hetaireias: 143-152.

 Archaelogikon Deltion

Samos XIX. Die Wasserleitung des Eupalinos auf Samos. Bonn: R. Habelt Vlg.

Kordellas, A. 1879. Αι Αθήναι Εξεταζόμεναι υπο Υδραυλικήν Έποψιν. Athens: Filokalia’s Publications.

Korres, M. (ed) 2010. Οι πρώτοι χάρτες της πόλεως των Αθηνών. Athens: Melissa Publications.

KTIMATOLOGIO S. A., http://gis.ktimanet.gr/wms/ktbasemap/default.aspx, Air photos of Greece

available online.

Lancaster, L. 2010. Parthian Influence on Vaulting in Roman Greece? An Inquiry into Technological

Exchange under Hadrian. American Journal of Archaeology 114: 447–72.

Landon, M. E. 2003. Beyond Peirene: Toward a Broader View of Corinthian Water Supply. In: C. K. Williams

II and N. Bookidis (eds) Corinth XX. Corinth, the Centenary, 1896–1996: 43-62. Princeton: American School

of Classical Studies at Athens.

               

dissertation, University of Pennsylvania.

Leigh, S. 2000. Interdisciplinary research on the aqueduct of Hadrian in the Athenian Agora. In: G. C.M.

Cura Aquarum in Sicilia. Proceedings of the Tenth International congress on the History of Water

management and Hydraulic Engineering in the Mediterranean Region (Syracuse, May 16-22, 1998): 117-124.

(BaBesch Supplement 6) Leiden: Peeters.

Leigh, S. 2001. A survey of the early Roman hydraulics in Athens, 65-82. In: A.O. Koloski-Osrtow (ed.) Water

use and hydraulics in the Roman city (Archaeological Institute of America, Boston, Colloquia and Conf.

Papers 3): 65-82. Dubuque (Iowa): Kendall/Hunt Publications.

Great Waterworks in Roman Greece

Leveau Ph. 2012. D’Alexandrie d’Égypte à Caesarea de Maurétanie: transfert de technologie hydraulique

et diffusion d’un nouveau modèle urbain. In: Stéphanie Guédon (ed.) Entre Afrique et Égypte: relations et

échanges entre les espaces au sud de la Méditerranée à l’époque romaine (Scripta Antiqua 49). Paris: Ausonius

Éditions.

Lightfoot, D. R. 2000. The Origin and Diffusion of Qanats in Arabia: New Evidence from the northern and

southern Peninsula. The Geographical Journal 166: 215-226.

Magee, P. 2007. Beyond the Desert and the Sown: Settlement Intensification in Late Prehistoric

Southeastern Arabia. Bulletin of the American Schools of Oriental Research 347: 83-105.

Liwa, Journal of

the National Center for Documentation and Research 3: 3-25.

Archeologikon Deltion

Lyktos 1: 66-99.



 Γ’ Διεθνές Συνέδριο Βοιωτικών Μελετών(Θήβα 4-8 Σεπτεμβρίου 1996)   



Praktikatēs en Athēnais Archaiologikēs Hetaireias 66:

132-142.

Mays, L.W. 2010. A Brief History of Water Technology during Antiquity: Before the Romans. In: L. W. Mays

(ed.) Ancient Water Technologies: 1-28. Arizona State University: Springer.

Papageorgiou-Venetas, A. 2010. Πόλεις και μνημεία στην Ελλάδα του Όθωνος. Athens: En Athinais

Archaeologike Etaireia.

Ύδρευσις των Αρχαίων Αθηνών. Athens: Eleutheri Skepsis Editions.

Paraskevopoulos, G. P. 1907. ΟιΔήμαρχοιτωνΑθηνών (1835-1907). Reprinted in 2001. Athens: Municipality

of Athens.

Papastavrou, E., 2006. Υπόγειοι λαξευτοί τάφοι τηςΑίγινας (Archaeologike Ephemeris 145). Athens.

Platonos-Yiota, M. 2004. Αχαρναί. Ιστορική και Τοπογραφική Επισκόπηση των Αρχαίων Αχαρνών, των

γειτονικών Δήμων και των οχυρώσεων της Πάρνηθας. Athens: Municipality of Acharnai.

Robinson, B. A. 2011. Histories of Peirene. A Corinthian fountain in three millennia. Athens: American School of

Classical Studies.

Salesse, E. 2001. Du nouveau à propos des galeries de captage émergentes. Quelques réflexions sur les

actes du séminaire du Collège de France de mars 2000. In: P. Briant (ed.) Irrigation et drainage dans

l’Antiquité. Qanâts et canalisations souterraines en Iran, en Egypte et en Grèce (Séminaire du Collège de

France, mars 2000). Topoi 11: 711-736.

Athenes et ses Environs. In: A. F. Stademann (ed.) Panorama von Athen. Munich.

Symeonoglou, S. 1985. The Topography of Thebes from the Bronze Age to Modern Times. Princeton: Princeton

University Press.

Tomlinson, R. A. 1969. Perachora: the remains outside the two sanctuaries. The Annual of the British School

at Athens 64: 155–258.

Pictorial Dictionary of Ancient Athens. Washington/ NY: Praeger Pub.

UAE 2006. Current status of aflaj in the Al Ain area. http://www.enhg.org/resources/articles/al_ain_

falaj/Al_%20Ain_Falaj_Report.pdf.

 Ορλάνδον

En Athinais Archaeologike Etaireia.

Voudouris, K. S., Christodoulakos, Y., Steiakakis, E. and Angelakis, A. N. 2013. Hydrogeological

Characteristics of Hellenic Aqueducts-Like Qanats. Water 5: 1326-1345.

Weingartner, H. 2007. Water supply by qanats. A contribution to water shortage in Mediterranean? In:

Proceedings of the 10th International ConferenceonEnvironmental ScienceandTechnology. Kos, Greece 5–7

September 2007: 1555-1561. Syros: University of the Aegean Department of Environmental Studies

Global Network for Environmental Science and Technology (Global-NEST).

Archaeological Landscapes of the Near East. University of Arizona Press.

Ancient

Iran from the Air: 12-27. Darmstadt: Zabern.

             

Rezakhani, K., De Schacht, T. 2012. From human niche construction to imperial power: long-term

trends in ancient Iranian water systems. Water History 4: 155-176.



             Oxford Handbook of

Engineering and Technology in the Classical World: 285-318. New York: Oxford University Press.

Zambas, C., Dounias, D. and Angistalis, G. 2017. The Aqueduct of Eupalinos on Samos, Greece, and its

Restoration. In: A. N. Angelakis, E. D. Chiotis, S. Eslamian and H. Weingartner (eds) Underground

Aqueducts Handbook: 63-80. Boca Raton: CRC Press.

         Mitteilungen des Deutschen

Archäologischen Institutes, Athenische Abteilung 2: 107-131.

Dr Eustathios D. Chiotis

Mining and Petroleum Engineer



The Institute of Geology and Mineral Exploration of Greece

echiotis@otenet.gr

Values and Costs in History: A Case Study on Estimating the Cost of Hadrianic Aqueduct’s Construction

Article

Full-text available

Apr 2022

The pursuit of wealth has been a basic occupation of humans; as it turns out, wealth increases life expectancy. Analyzing global data, we show that money, probably connected with medical care, increase life expectancy. However, the base of real wealth is access to the Water–Energy–Food nexus, and the access to this also increases life expectancy. The first objective of this study was to compare the present values of wealth with antiquity, and we showed that about 1.4 billion people live in the present under the average lower wages of antiquity. As a case study, we analyze the construction of the Hadrianic aqueduct. We present a detailed description of the construction and the used methods, and we identify the total requirement of labor–time. Then, we investigate the wages of various occupations in the first century AD. The second objective of this study was the estimation of the total cost of daily wages for the construction of the project and the effect of the aqueduct on Athenians’ quality of life. Finally, we show that, today, about two billion people live with less available water than Athenians had with the Hadrianic aqueduct in the second century A.D.

Issues of prosperity: Stochastic evaluation of data related to environment, infrastructures, economy and society

Thesis

Full-text available

Jun 2022

G.-Fivos Sargentis

Living organisms pass through life seeking prosperity in a materialistic world. There are different meanings of prosperity. Some people think that it is measured in money, others relate it to pleasure and life satisfaction, while others link it to spirituality. However, it could be argued that the basic human needs related to the Water, Energy and Food (WEF) compose a nexus not only necessary for the survival of humans, but able to explain their prosperity as well. Unfortunately, decision-making in modern world is largely driven by economic aspects and monetarist policies. Koutsoyiannis (personal communication) notes that water, energy and food are not derived by money; rather money and economic growth derives from the availability and the access to water, energy and food. In this thesis, we study critical issues of prosperity rationally, using publicly available data, historical evidences and stochastic tools. The studied issues are based on the WEF nexus but extend to various other societal, environmental and cultural aspects of human life in societies, ranging from social stratification and urban clustering, to the aesthetic quality of surrounding environment.

THE ROMAN AQUEDUCT OF KNOSSOS, A MODEL FOR NINETEENTH-CENTURY AQUEDUCT DESIGN

Article

Full-text available

Feb 2023

In this paper, we present findings from a field inspection of the Knossos aqueduct undertaken in 2019. A key contribution of our fieldwork was the architectural identification of the Roman channel underlying the nineteenth-century wall of the Ottoman-Egyptian aqueduct supplying Iraklio. While reuse of the Roman aqueduct in the nineteenth century was known from historical reports, the structural overlap has never been identified in the field or documented archaeologically, until now. We recorded the Roman channel lined with opus signinum running along the base of the nineteenth-century aqueduct's wall between Fundana and Spilia. Through this realisation in the field, we were able to establish diagnostic styles of masonry for both periods. Our architectural distinction between the overlaid aqueducts allowed us to integrate previously disarticulated components of the later system, like the reused Roman tunnel at Skalani and the nineteenth-century bridge at Spilia, into an integrated Ottoman-Egyptian water supply for Iraklio. As we approached Knossos from Spilia, we were also able to identify the point at which the Venetian aqueduct supplying Iraklio converged with the Roman system. Consequently, our 2019 fieldwork not only mapped the length of the Roman aqueduct supplying the city of Knossos but also that section of the nineteenth-century Ottoman-Egyptian aqueduct of Iraklio built directly over it and a shorter tract of the Venetian aqueduct of Iraklio that either ran alongside it or was, in turn, itself partially overlaid by the nineteenth-century system.

The Roman aqueduct of Philippi

Article

Jan 2019

Colonia Iulia victrix Philippiensium was founded by Mark Antony after the battle of Philippi in 42 BC and was re-founded by Octavian as colonia Iulia Philippiensis, then Augusta, after the naval battle in Actium in 31 BC. Philippi was one of the most important stations along the Egnatia Road. During the Antonine period, in the second half of the 2nd c. AD, the city’s center was reconstructed in its most monumental form. The aqueduct for the water supply of the city has been dated to the same period. It is an open-channel aqueduct with a barrel vaulted cover that collected water from the Kefalari area (karstic springs of Voirani), about seven kilometers north-northwest of Philippi.

The cisterns of the Bisti promontory at Hermione. With a preliminary description of the Roman aqueduct

Article

Nov 2021

Patrik Klingborg

This article reports the findings of the fieldwork exploring the cisterns at the Bisti promontory of Hermione, executed as part of a collaboration between the Ephorate of Antiquities of the Argolid and the Swedish Institute at Athens. In order to better understand the function of the cisterns within the water supply system, the article begins by presenting an overview of existing water resources in the area, primarily the naturally occurring sources and the city’s 2nd-century AD Roman aqueduct. Following this the study describes the remains of the 14 potential cisterns on the Bisti. Based on the empirical material the similarities and contrasts between these are explored, as well as what they can tell us about the history and life in ancient Hermione. In particular, the article suggests that the presence of the cisterns contributes to our understanding of the urban fabric of the city, and reveals important information about when the city was moved from the Bisti to the nearby Pron Hill.

The Shaft Tunnel of the Biar aqueduct of Jerusalem: Architecture, hydrology, and dating

Article

May 2021

The Biar aqueduct is the most sophisticated of the aqueducts supplying water to Jerusalem during classical periods. It includes the Biar spring, an underground Shaft Tunnel, ∼3 km long harvesting groundwater (cuniculus), a dam, a surface channel, and a tunnel traversing a ridge. We conducted a survey of the Biar underground Shaft Tunnel, mapping all of its accessible parts (∼1200 m), facilitating a new understanding of the cutting‐edge hydrogeologic and engineering skills used for this project. The last ∼536 m of the Biar Shaft Tunnel was constructed with a unique ashlar channel, divided into segments, designed to withstand different loads: (a) on a mechanically weak bedrock, a channel with arched gables and barrel vault specus was built within a hewn winding tunnel, (b) when dug as an open shallow trench, a channel roofed with complex gables of ashlars with drafted margins was built, and (c) to release hydraulic pressure, a channel roofed with alternations of barrel vaults and simple gables set perpendicular to the course of the tunnel was constructed. 14C dating of plaster samples suggests that the Biar aqueduct was built in the mid‐first century CE and reconstructed during Aelia Capitolina in the second century CE.

Integrating geology into archaeology: the water supply of Piraeus in Antiquity

Article

Full-text available

Jan 2019

Eustathios Chiotis

The peninsula of Piraeus is composed of the Mounichia Hill and the rocky Akti, connected through the NE-SW trending isthmus between the Kantharos and Zea harbors (Figure 1a). The great geomorphological advantage of the peninsula is its natural and safe harbors – Kantharos with the innermost Kophos Port, Zea and Mounichia – at a reasonable distance of eight kilometers from Athens.

Updated appraisal of ancient underground aqueducts in Greece

Chapter

Full-text available

Nov 2016

Qanat and falaj: Polycentric and multi-period innovations Iran and the United Arab Emirates as case studies

Chapter

Full-text available

Nov 2016

Rémy Boucharlat

This chapter deals with the ancient underground aqueducts in the United Arab Emirates (UAE) and Iran. Some regional cases in other countries will be occasionally mentioned when dealing with the questions of the chronology and diffusion of the earliest witnesses. All the documents and case studies presented here can be grouped under the generic term shafts-and-gallery aqueducts (as sug- gested by E. Chiotis in his introduction Chapter 1 in this volume, tapping groundwater). The last words of that de nition aim at excluding the (partly or totally) subterranean aqueducts collecting water from springs, lakes, and dams. Conversely, the depth of the mother well cannot be used for making a distinction between the so-called Iranian qanāt and the other types of qanāt or falaj, as has been often assumed in hundreds of papers. As a matter of fact, the vocabulary used by the local people and hence by scholars in several countries, and the word qanāt, from Iran to Egypt via Syria, 279 280 Underground Aqueducts Handbook is limited, and one cannot reduce this term to the aqueduct with a deep mother well tapping a deep aquifer.* Therefore, qanāt for Iran and falaj for the UAE and the Sultanate of Oman are employed here in the broad meaning as de ned above. Both of them encompass a large “family” of water catchments. Leaving aside any hydrogeological aspect, the viewpoint of this chapter is rst the technical skills and knowledge of the population digging underground galleries, as seen by an archaeologist. Therefore, the aims are: (1) to present the earliest evidence of underground galleries in the early rst millennium BC in South-East Arabia; (2) to mention some mid- rst millennium BC isolated cases, in Egypt and Iran, both in a very peculiar geological context; (3) to stress the lack of direct archaeological evidence in Iran and even of indirect evidence until the mid- rst millennium AD; therefore, one cannot infer an Iranian in uence on the other side of the Persian Gulf in antiquity; and (4) to assume from the written evidence the existence of qanāt tapping deep aquifers in Iran very likely in the mid- rst millennium AD. The hypothesis of radial diffusion from a unique center does not match with the archaeological evidence. The qanāt and falaj in their general meaning may well be a polycentric innovation in different geographical contexts at different periods, especially for the rst period, the rst mil- lennium BC. For that period, there are generally short galleries capturing water from the shallow aquifer formed by seepage from the watercourse of a wadi. In the early rst millennium AD, some countries continued (Egypt and Libya) or developed (Syria, Northern Iraq) rather simple and shallow underground galleries. At the same time, very likely in Iran, the improvement of knowledge, may be science and techniques, allowed the specialists to detect unsuspected and deeper water resources. These galleries of the “second generation” are well attested in the writ- ten sources in Iran and have been illustrated by thousands of qanāt from that time to nowadays, while shallow galleries are still in use. For this second “generation,” a diffusionist model may be reconstructed.

Underground aqueducts types, definitions, and general conclusions

Chapter

Full-text available

Nov 2016

Eustathios Chiotis

Geological aspects on the sustainability of ancient aqueducts of Athens

Article

Full-text available

Dec 2016

Geological aspects of the ancient aqueducts of Athens are examined with particular emphasis on the hydrogeological and geotechnical conditions which made possible their continuous function throughout the centuries up today. In particular, the sustainability of the Hymettos and the Hadrianic aqueducts is investigated and attributed to the capture of underground water and the skillful construction of the tunnels. The geometry of the Hadrianic aqueduct and the situation in the ancient tunnel are presented based on data obtained during the resumption of the ancient work in the 19th and 20th centuries. The up to forty meters deep wells along the tunnel continue to collect underground water at many places along the twenty kilometers long path of the Hadrianic aqueduct. Similarly, the Hymettos aqueduct still irrigates the National Garden with a daily rate of about one thousand cubic meters. The stability of the ancient tunnels is correlated with the surrounding geological formations; it turns out that the stability problems of the ancient tunnels are comparable to those met in the tunneling works of the Metro lines, particularly in the Athenian Schists. A new part of an ancient aqueduct was crossed during the Metro works which is considered to belong to the Hymettos aqueduct.

Water supply of Athens in Antiquity

Chapter

Full-text available

Jan 2012

The Snake figuration in Iron Age Society

Article

Jan 2011

Qanat Fir'aun: An underground roman water system in Syria and Jordan

Chapter

Nov 2016

M. Döring

The aqueduct of eupalinos on Samos, Greece, and its restoration

Chapter

Nov 2016

Chronique des fouilles

Article

Jan 2011

Nicolaos Papachatzis

Parthian Influence on Vaulting in Roman Greece? An Inquiry into Technological Exchange Under Hadrian

Article

Apr 2010

Lynne Lancaster

The earliest known Roman examples of "pitched-brick" vaulting appear in hydraulic systems in the first half of the second century C.E. at Athens, Eleusis, and Argos, while others at Dion and Gortyn may also date from the same period. I argue, based on an analysis of the constructional details, that the technique was introduced from Parthia by military construction experts specializing in hydraulics, who had been involved in Trajan's Parthian War and, upon returning to home bases in the West, brought new construction techniques. The examples from Athens occur in the Hadrianic aqueduct and thus can be securely dated to the period just after the war. Those from Eleusis, which also relate to an aqueduct project, can likewise be dated to the reign of Hadrian or Antoninus Pius. The most impressive example occurs at the so-called Serapeum (Bath A) at Argos, which is among the largest known barrel vaults using this technique. It has been dated by the excavators to ca. 100 C.E.; however, I argue that this structure was originally built somewhat later, under Hadrian, and that it was part of a larger project at Argos.

The Hadrianic aqueduct of Athens and the underlying tradition of hydraulic engineering

Recommended publications

Comparison between two methods for quickly determining hydraulic parameters of unsaturated soils

Flow conditions in the round pool pass

Underground aqueducts types, definitions, and general conclusions

Underground aqueducts handbook

Eustathios D. Chiotis, Minoan hydraulic tradition and technology transfer to Thebes and Corinth in G...

The International Journal for the History of Engineering & Technology  0 Views 0 CrossRef citations...